CN107463166A - Automatic running device and its control traveling method - Google Patents
Automatic running device and its control traveling method Download PDFInfo
- Publication number
- CN107463166A CN107463166A CN201610389387.3A CN201610389387A CN107463166A CN 107463166 A CN107463166 A CN 107463166A CN 201610389387 A CN201610389387 A CN 201610389387A CN 107463166 A CN107463166 A CN 107463166A
- Authority
- CN
- China
- Prior art keywords
- cell
- result
- region
- subimage block
- recognition result
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000004040 coloring Methods 0.000 claims abstract description 13
- 230000008859 change Effects 0.000 claims description 14
- 235000013399 edible fruits Nutrition 0.000 claims description 5
- 238000012423 maintenance Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 10
- 230000009471 action Effects 0.000 description 8
- 239000006185 dispersion Substances 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 230000009286 beneficial effect Effects 0.000 description 4
- 239000000428 dust Substances 0.000 description 4
- 238000005520 cutting process Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000002156 mixing Methods 0.000 description 3
- 241000208340 Araliaceae Species 0.000 description 2
- 244000025254 Cannabis sativa Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 239000003973 paint Substances 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000005859 cell recognition Effects 0.000 description 1
- 239000004568 cement Substances 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000009931 harmful effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of method for controlling automatic running device walking, it is characterised in that comprises the following steps：S10, the image for obtaining the automatic running device steps target region；S20, several cells are divided the image into, each cell has at least one adjacent cells lattice；S30, the colouring information of specified pixel in the cell and the texture eigenvalue of the cell identify whether target area corresponding to the cell is working region, and obtain recognition result；S40, divide the image into several subimage blocks, each subimage block includes several adjacent cells, whether target area is walkable region according to corresponding to the recognition result of the cell in the subimage block judges the subimage block, and obtains judged result；S50, according to judged result, control the direction of travel of automatic running device.
Description
Technical field
The present invention relates to a kind of automatic running device, and the method for controlling the automatic running device to walk.
Background technology
With the continuous progress of computer technology and artificial intelligence technology, similar to the automatic row of intelligent robot
Walk equipment and have begun to the life for coming into people slowly.The companies such as Samsung, Electrolux develop it is complete from
Simultaneously market has been put into dynamic dust catcher.This usual compact of automatic dust catcher, is integrated with environmentally sensitive
Device, from drive system, dust collecting system, battery and charging system, can be without manual control, voluntarily indoors
Cruise, the auto-returned bus stop when energy is low, docks and charges, and then proceedes to dust suction of cruising.Meanwhile
The companies such as Haas Ke Wana develop similar intelligent grass-removing, its can automatically in the lawn of user mow,
Charging, interfere without user.Just without putting into energy again after once being set due to this automatic mowing system
Management, freed during the housework that user is uninteresting from cleaning, lawn maintenance etc. and wastes time and energy is worked, because
This is greatly welcome.
The travel region of existing automatic mower is generally by the boundary line for setting physics, such as wire or hedge
Basketry, the boundary line of automatic mower detecting physics is to determine travel region.The process of border wiring is cumbersome,
Time and effort consuming, and non-grass areas is likely present in boundary line, or also presence needs to mow outside boundary line
Region, it is dumb, inconvenient using the method for physical boundary line.
And the solution method of travel region is identified and determined using electronically, it is various due to travel region
Property, noise is often produced using existing means, identifies that the accuracy of travel region is very low, so as to influence certainly
The judgement of dynamic walking arrangement, it is easy to travel region is left, so as to be made to the normal work of automatic running device
Influenceed into certain.
It is therefore desirable to prior art means are improved so that travel region is more accurately identified, so as to
Automatic running device is facilitated to be operated.
The content of the invention
In view of this, an object of the present invention is that providing one kind accurately identifies target area and can control certainly
The method and the automatic running device of application this method that dynamic walking arrangement is accordingly walked according to recognition result.
To achieve the above object, the technical solution adopted in the present invention is：One kind control automatic running device row
The method walked, it is characterised in that comprise the following steps：S10, obtain the automatic running device steps target
The image in region；S20, several cells are divided the image into, each cell has at least one phase
Adjacent cell；S30, the colouring information of specified pixel in the cell and the cell
Texture eigenvalue identifies whether target area corresponding to the cell is working region, and obtains recognition result；
S40, several subimage blocks being divided the image into, each subimage block includes several adjacent cells,
The target area according to corresponding to the recognition result of the cell in the subimage block judges the subimage block is
No is walkable region, and obtains judged result；S50, according to judged result, control automatic running device
Direction of travel.
Preferably, after methods described is additionally included in acquisition recognition result, for each cell, according to described
The recognition result of cell and its adjacent cells lattice, adjust the recognition result of the cell.
Preferably, methods described further comprises the steps：S61, a cell is selected, and obtain it
Recognition result；S62, statistics are with the quantity with the adjacent cells lattice of same identification result in step S61；S63、
Ratio of the quantity in total adjacent cells lattice quantity in calculation procedure S62；If S64, the ratio exceed or
Reach preset value, then the recognition result for the cell that maintenance step S61 is specified is constant；If the ratio is less than pre-
If value, then change the recognition result of cell that step S61 is specified, wherein the 4th preset value be more than or
Equal to 50%.
Preferably, the adjacent cells lattice include list adjacent with the selected cell in transverse direction and longitudinal direction
First lattice.
Preferably, the adjacent cells lattice also include with the transverse direction and longitudinal direction on the direction of 45 degree angles and
The adjacent cell of the selected cell.
Preferably, methods described further comprises the steps：S66, a cell is selected, described in acquisition
For cell for the reliability Y1 of its recognition result, reliability Y1 is a numerical value between 0~100%；S67、
1-Y1 is calculated, and is N1 by result queue；S68, all adjacent cells lattice pair for obtaining the selected cell
In reliability Ya, Yb ... of its recognition result, reliability Ya, Yb ... is a numerical value between 0~100%；
S69,1-Ya, 1-Yb ... are calculated, and be Na, Nb ... result queue；S70, the weighting Ya, Yb ...
Summation obtains weighted sum Y2, and it is N2 that the weighted sum Nb, Nc ..., which obtains weighted sum, wherein, weighting system
Number all same；S71, Y1+ α N1 and Y2+ α N2 result being calculated respectively and compares its size, wherein α is
Coefficient；If S72, Y1+ α N1 result are more than or equal to Y2+ α N2 result, remain described and specify
The recognition result of cell is constant, if result of the Y1+ α N1 result less than Y2+ α N2, described in change
The recognition result for the cell specified.
Preferably, the S40 steps further comprise the steps：S41, described image is divided into it is some
Individual subimage block, the quantity for the cell that each subimage block includes is obtained, is marked as B；Collect institute
The recognition result of the cell in subimage block is stated, statistics recognition result is the quantity of the cell of working region,
And it is marked as A；If A：B is less than the 3rd preset value, then judges target area corresponding to the subimage block
Domain is not walkable region, otherwise, judges that walkable region is in target area corresponding to the subimage block.
Preferably, methods described also includes same target area is continuously shot to form multiple image, according to every
In two field picture the judged result of same subimage block judge target area corresponding to the subimage block whether be can
Travel region, and obtain judged result.
Preferably, methods described further comprises the steps：S81, shape is continuously shot to same target area
Into multiple image；A subimage block in S82, a selected wherein two field picture, is sentenced by step S40
Disconnected result；S83, initial parameter value is set, and parameter value entered according to the step S82 judged results obtained
Row computing, if judged result is walkable region, increase and judged result pass in the initial parameter value
First parameter of connection turns into current parameter value；If judged result is not walkable region, the parameter is kept
It is worth constant；S84, next two field picture is selected, and the judged result obtained according to step S82 is to current parameter
Value carries out computing, if judged result is walkable region, increase and judgement on the current parameter value
As a result the first parameter associated turns into new current parameter value；If judged result is not walkable region, protect
It is constant to hold the current parameter value；The size of the current parameter value of S85, comparison and threshold value, if current ginseng
Numerical value is more than or equal to threshold value, then assert that target area corresponding to the subimage block is walkable region.
Preferably, the step S84 is still further comprised：To current ginseng after next two field picture is selected
Before numerical value carries out computing, current parameter value subtracts default second parameter, and second parameter is less than
First parameter.
Preferably, the subimage block includes three middle part, left part and right part subimage blocks, corresponds to mesh respectively
Mark intermediate region, left field and the right side area in region.
To achieve the above object, another technical scheme of the present invention is：A kind of automatic running device,
It is characterized in that：Image collecting device including housing, on housing, described image harvester are used for
Photographic subjects region simultaneously generates image, the walking module for driving the automatic running device to walk, described in connection
Image collecting device and walking module to control the main control module that automatic running device works, wherein, the master
Control module includes division unit, recognition unit, judging unit and control unit, and the division unit is described in
Image is divided into several cells, and division result is passed to the recognition unit, the recognition unit
Identify whether target area corresponding to the cell is working region, and recognition result is passed to judgement list
Member, judging unit judge whether the subimage block corresponding region comprising some cells is walkable region, and
Judged result is passed to control unit, described control unit controls the row of walking module according to judged result
Walk direction.
Preferably, the main control module also includes amending unit, and the amending unit is directed to each cell,
According to cell and its recognition result of adjacent cells lattice, the recognition result of the cell is adjusted.
Preferably, the adjacent cells lattice include list adjacent with the selected cell in transverse direction and longitudinal direction
First lattice.
Preferably, the adjacent cells lattice also include with the transverse direction and longitudinal direction on the direction of 45 degree angles and
The adjacent cell of the selected cell.
Preferably, the judging unit also includes subimage block division unit, the subimage block division unit
Divide the image into several subimage blocks, the knowledge for the cell that judging unit includes according to the subimage block
Other result judges whether corresponding subimage block is walkable region.
Preferably, the subimage block includes three middle part, left part and right part subimage blocks.
Preferably, the main control module also includes the recording unit that record has initial parameter value, and described image is adopted
Acquisition means are continuously shot to form multiple image to same target area, and the judging unit is in each two field picture
Same subimage block is judged and obtains judged result, recording unit is according to its judged result to parameter value
Computing is carried out, when parameter value is more than or equal to threshold value, then assert target area corresponding to the subimage block
It is walkable region.
Compared with prior art, beneficial effects of the present invention are：Unit is divided into for the image of target area
Lattice, it is microcosmic on each cell is identified, macroscopically integrate multiple cells recognition result carry out
It is comprehensive to distinguish, so as to lift the accuracy of identification object region, beneficial to automatic running device in target area more
Add accurate walking.
Another object of the present invention is to provide a kind of method for accurately identifying target area and using this method
Automatic running device.
To achieve the above object, a kind of technical scheme of the present invention is：One kind is set on automatically walk
The recognition methods of the target area of standby walking, it is characterised in that the recognition methods comprises the following steps：S10、
Obtain the image on the automatic running device steps target region；S20, divide the image into several
Cell, each cell have at least one adjacent cells lattice；S30, the finger in the cell
The colouring information of fixation element and the texture eigenvalue of the cell identify target area corresponding to the cell
Whether domain is working region, and obtains recognition result；S60, for each cell, according to adjacent cells
The recognition result of lattice, change or maintain the recognition result obtained by step S30.
Preferably, the S60 steps further comprise the steps：S61, a cell is specified, and
Obtain its recognition result；S62, statistics have and the adjacent cells lattice of same identification result in step S61
Quantity；The ratio of quantity in S63, calculation procedure S62 in total adjacent cells lattice quantity；If S64, it is somebody's turn to do
Ratio exceedes or reached the 4th preset value, then the recognition result for the cell that maintenance step S61 is specified is constant；
If the ratio is less than the 4th preset value, change the recognition result for the cell that step S61 is specified, wherein institute
The 4th preset value is stated more than or equal to 50%；S65, above-mentioned S61~S64 steps are performed to all cells.
Preferably, the adjacent cells lattice include cell adjacent with the cell in transverse direction and longitudinal direction.
Preferably, the adjacent cells lattice also include with the transverse direction and longitudinal direction on the direction of 45 degree angles and
The adjacent cell of the cell.
Preferably, the S60 steps further comprise the steps：S66, a cell is specified, obtained
It is one between 0~100% to take the cell reliability Y1 for its recognition result, reliability Y1
Numerical value；S67,1-Y1 is calculated, and be N1 by result queue；S68, obtain all of the designating unit lattice
For adjacent cells lattice for reliability Ya, Yb ... of its recognition result, reliability Ya, Yb ... is 0~100%
Between a numerical value；S69,1-Ya, 1-Yb ... are calculated, and be Na, Nb ... result queue；S70、
Ya, Yb ..., weighted sum obtains weighted sum Y2, and the weighted sum Nb, Nc ... obtains weighted sum N2,
Wherein, weight coefficient all same here；S71, calculate Y1+ α N1 and Y2+ α N2 result simultaneously respectively
Compare its size, wherein α is coefficient, if Y1+ α N1 are more than or equal to Y2+ α N2, maintain described in
The recognition result for the cell specified is constant, if Y1+ α N1 result is less than Y2+ α N2 result,
Change the recognition result of the cell specified；S72, all cells are performed with above-mentioned S66~S71 steps
Suddenly, until the recognition result of all cells no longer changes.
To achieve the above object, another technical scheme of the present invention is：A kind of automatically walk is set
It is standby, it is characterised in that：Image collecting device including housing, on housing, described image harvester
For photographic subjects region and image is generated, drives the walking module of the automatic running device walking, connection
Described image harvester and walking module to control the main control module that automatic running device works, wherein, institute
Main control module division unit, recognition unit and amending unit are stated, the division unit is divided into described image
Several cells, the recognition unit identify whether target area corresponding to the cell is working region,
And recognition result is passed to amending unit, the amending unit is directed to each cell, according to adjacent cells
The recognition result of lattice, change or maintain the recognition result obtained by recognition unit.
Preferably, the cell adjacent with the cell includes adjacent with the cell in transverse direction and longitudinal direction
Cell.
Preferably, it is in 45 degree of angles that the cell adjacent with the cell, which also includes with the transverse direction and longitudinal direction,
Direction on the cell adjacent with the cell.
Compared with prior art, beneficial effects of the present invention are：Unit is divided into for the image of target area
Lattice, it is microcosmic on each cell is identified, macroscopically integrate multiple cells recognition result carry out
Amendment, so as to lift the accuracy of identification object region.
Brief description of the drawings
Purpose, technical scheme and the beneficial effect of invention described above can be real by following
The detailed description of the specific embodiment of the existing present invention, describes in combination with accompanying drawing and clearly obtains.
Identical label and symbol in accompanying drawing and specification are used to represent identical or equivalent element.
Fig. 1 is the schematic diagram that the automatic running device of the embodiment of the present invention is walked in target area.
Fig. 2 is the schematic diagram in the automatic running device photographic subjects region in Fig. 1.
Fig. 3 is the schematic diagram of Fig. 1 automatic running device division target area.
Fig. 4 is the schematic diagram of each several part module of Fig. 1 automatic running device.
Fig. 5 is the schematic flow sheet of the control automatic running device traveling method of the embodiment of the present invention.
Fig. 6 is the detailed of the step S60 between step S30 and step S40 in one embodiment of the invention
Thin schematic flow sheet.
Fig. 7 is the step S60 between step S30 and step S40 in another embodiment of the present invention
Detailed process schematic diagram.
Fig. 8 is the detailed process schematic diagrames of step S40 in one embodiment in Fig. 5.
Fig. 9 is the detailed of the step S80 between step S40 and step S50 in one embodiment of the invention
Thin schematic flow sheet.
1st, automatic running device 2, image collecting device 3, main control module
4th, walking module 5, operational module 6, energy module
9th, driving wheel 10, housing 11, auxiliary wheel
12nd, division unit 13, color extraction unit 14, computing unit
15th, comparing unit 16, memory cell 17, texture blending unit
18th, texture comparing unit 19, recognition unit 20, information extraction unit
21st, information changes unit 22, judging unit 23, subimage block division unit
28th, target area 32, amending unit 33, recording unit
50th, working region 51, the isolated island of non-active area 52
Embodiment
Presently preferred embodiments of the present invention is described in detail below in conjunction with the accompanying drawings so that advantages of the present invention and
Feature can be easier to be readily appreciated by one skilled in the art, apparent so as to be made to protection scope of the present invention
Clearly define.
Fig. 1 show the schematic diagram that the automatic running device of one embodiment of the invention is walked in target area.From
Dynamic walking arrangement 1 can automatically walk on ground or other working surfaces, can also while walking
It is operated.Automatic running device 1 can be automatic cleaner, automatic mower, automatic pruning machine etc..
In the present embodiment, automatic running device is automatic mower.Different according to the object of work, ground can be with
It is divided into working region 50 and non-active area 51.Working region 50 refers to that user wants to allow automatic running device
The region that walking is passed through and is operated, and non-active area 51 refers to that user does not want to allow automatic running device to pass through
The region crossed.In the present embodiment, because automatic running device is automatic mower, its work is cut for execution
Grass.Therefore travel region 50 can with but be not limited to meadow, and non-active area 51 can with but be not limited to
Cement road, great Shu, pond, fence, timber, corner etc..Under normal circumstances, meadow is all blocking in flakes
Formed, rather than travel region can be located at around meadow, can also be surrounded by meadow and form isolated island 52,
So isolated island 52 and a kind of form of expression of non-travel region.In the present invention, the He of non-active area 51
The intersection of working region 50 can be not provided with boundary line, and automatic running device 1 utilizes the He of working region 50
The difference of non-active area 51 visually is identified.
With reference to Fig. 2 and Fig. 3, automatic running device 1 has the image of housing 10 and installation on the housing 10
Harvester 2.Image collecting device 2 shoots the image of the front region of automatic running device 1.Positioned at automatic
The ground region in the front of walking arrangement 1 is the target area 28 of automatic running device walking.Target area 28
It is likely to be working region, it is also possible to non-active area, it is also possible to travel region and non-travel region
Set.And automatic running device 1 is in order to performing the purpose normally walked in travel region, it is necessary to working as
Preceding target area 28 is identified.Therefore automatic running device 1 can be to this using image collecting device 2
Target area 28 shoots and forms the image on target area 28.Therefore the side of automatic running device is controlled
Method includes step S10, that is, generates the image on the steps target region of automatic running device 1.In this reality
Apply in example, the viewfinder range of image collecting device 2 is a FX, and 90 degree of the angular field of view such as fixed is extremely
120 degree.Viewfinder range can also be movable in other alternative embodiments, can choose one in angular field of view
Angular range is determined as actual viewfinder range, such as chooses 90 degree of scopes for being located at middle part in 120 degree of angular field of view
As actual viewfinder range.The information of target area, such as the hypsography of target area are included in the image
Situation, distribution of color situation, texture situation etc..
Refer to Fig. 4, in addition to image collecting device 2, automatic running device 1 also include main control module 3,
Walking module 4, operational module 5 and energy module 6.Main control module 3 respectively with walking module 4, Working mould
Block 5, energy module 6 and image collecting device 2 are electrical connected, and play control automatic running device 1
The effect of work.
Walking module 4 includes wheel group and the running motor for drive wheel group.Wheel group can have a variety of setting sides
Method.Usual wheel group includes the auxiliary wheel 11 of the driving wheel 9 and Auxiliary support housing 10 by running motor driving,
The quantity of driving wheel 9 can be for 1,2 or more.As shown in Fig. 2 with automatic running device 1
Moving direction be used as front side, the side relative with front side is rear side, and the both sides adjacent with front and rear side are respectively
The left and right sides.In the present embodiment, the driving wheel 9 of automatic running device 1 is 2, respectively positioned at a left side
The revolver 91 of side and the right wheel 92 positioned at right side.Revolver 91 and right wheel 92 are on automatic running device 1
Axis is symmetrical arranged.Revolver 91 and right wheel 92 are preferably located at the rear portion of housing 10, auxiliary wheel 11
In front portion, setting can also be replaced in other embodiments certainly.
In the present embodiment, a drive motor is each coupled in revolver 91 and right wheel 92, to realize that differential is defeated
Go out and turned to controlling, so as to the purpose for reaching to turn left or turning right.Revolver 91 and right wheel 92 also being capable of constant speed
Output, so as to reach the purpose advanced or retreated.Drive motor can be directly connected to driving wheel, but also may be used
To set transmission device between drive motor and driving wheel 9, such as common planetary gear train in the art.
In other examples, driving wheel 2 is may also set up, drive motor 1, in this case, is driven
Motor passes through the second actuator drives right wheel 92 by the first actuator drives revolver 91.It is i.e. same
Motor passes through different actuator drives revolver 91 and right wheel 92.
Operational module 5 is used to perform specific work.In the present embodiment, operational module 5 is specially blanking punch
Block, including cutting part (not shown) and the cutting motor (not shown) of driving cutting part for mowing.
Energy module 6 is used to provide energy to the operation of automatic running device 1.The energy of energy module 6 can
Think gasoline, battery bag etc., energy module 6 is included in the chargeable of the interior setting of housing 2 in the present embodiment
Battery bag.When work, battery bag discharges electric energy to maintain automatic running device 1 to work.In non-work
When work, battery may be coupled to external power source with electric energy supplement.Especially, set for more humane
Meter, when detecting the not enough power supply of battery, the searching charging bus stop that automatic running device 1 can voluntarily (is schemed
Do not show) electric energy supplement.
As shown in figure 3, image collecting device 2 passes to master after obtaining the image on target area 28
Control module 3.Main control module 3 includes division unit 12.Division unit 12 is used to image divide several lists
First lattice.All cells are combined into an image, and each cell occupies a part for whole image.Cause
This each cell contains the identification information of the parts of images.The size of each cell is basically identical.Separately
Outside, several cells form matrix array.The matrix array is respectively transversely and Longitudinal extending.In horizontal stroke
Upwards, about 20 cells are arranged in a row；And in the vertical, about 20 cells are arranged in a row.
In various embodiments, the cell quantity of transverse direction and longitudinal direction arrangement can be with inconsistent.And each cell
All there is at least one cell adjacent thereto.For the cell positioned at array intermediate region, Mei Gedan
First lattice have is located at up and down totally 4 cells adjacent thereto, in other words, 4 units respectively
Lattice are respectively in the horizontal or longitudinal upper adjacent with that cell.Certainly, the adjacent meaning is also not necessarily limited to
4 directions in lower left and right.In a further embodiment, for the cell have up and down and it is upper it is left,
Eight upper right, bottom left, bottom right directions totally 8 cells adjacent thereto, in other words, except in transverse direction
And it is longitudinally adjacent outside, with can also be adjacent with cell on direction of the transverse direction and longitudinal direction in 45 degree of angles.
And may be without 4 units adjacent thereto for the cell positioned at array edges region, each cell
Lattice, but at least have one it is adjacent thereto.Therefore the method for control automatic running device further comprises step
S20, i.e. image are divided into several cells, and each cell is adjacent with least one other cell.
After the division unit lattice of division unit 12, whether target area corresponding to beginning recognition unit lattice is capable
Walk region.Idiographic flow method is as follows：Main control module 3 first reads the identification information that unit lattice are included.
In the present embodiment, the identification information that cell is included includes colouring information and texture information.At other
In embodiment, the information that cell includes can be colouring information and other kinds of information.Due to unit
Lattice are a parts for image, and image includes the information to target area.Therefore cell necessarily contains
The information of corresponding target area, also include colouring information certainly.Help to sentence by reading the identification information
Target area corresponding to the cell of breaking is management work region or non-active area.Due to as work
The meadow in region is green, and the road, soil as non-active area are not then green, if therefore
The colouring information for identifying the cell is green, then it is believed that being travel region corresponding to the cell.If
It is not green to identify colouring information, then it is believed that being non-travel region corresponding to the cell.Of course for
Accuracy is further lifted, the non-travel region of some situations is also green, such as some are artificial treated
Body surface painted green paint, in this case the color of travel region and non-travel region all for green,
Differentiated from colouring information, be not easy to distinguish travel region and non-travel region.Therefore also need to add
Identification to texture information.Because in the case of being also green in that non-travel region, generally it has rule
Texture then, although and the meadow of travel region also for green, it is careless must grow be not so have rule
Then, therefore its texture is irregular.And then if identify the colouring information of the cell as green, and
Crossgrained, then it can assert corresponding to the cell it is travel region.If color is not green or texture
Rule, then can assert corresponding to the cell it is non-travel region.Certainly, in other embodiments, also may be used
It is just detailed not to the utmost herein to reach the purpose of examination travel region and non-travel region by identifying other information
State.
On the other hand, main control module 3 is gone back color extraction unit 13, computing unit 14, comparing unit 15 and deposited
Storage unit 16.Then main control module 3 is believed colouring information with default by the colouring information of extraction unit lattice
Whether breath is compared, be that travel region is identified to cell according to comparative result.Specific method is such as
Under：Many pixel cells are contained because each cell is actual, and the color that pixel cell is shown is only
One.Therefore the effect of color extraction unit 13 is the color of each pixel cell in extraction unit lattice, especially
Ground, extraction is three primary colors (RGB) component.And presupposed information refers to that default is made with reference to comparison other
Information.In the present embodiment, presupposed information refers to the numerical value model for the three primary color components for having predetermined color
Enclose.In the present embodiment, predetermined color refers to green.Compare the three primary color components and predetermined color of a pixel
Three primary color components, if the three primary color components of a pixel respectively fall in the number of the three primary color components of predetermined color
It is worth scope, then judges the color of the pixel for predetermined color.If illustrating not falling within the number range,
Judge the color of the pixel for non-predetermined color.In a further embodiment, memory cell 16 has predetermined face
Default tone value (Hue) scope of color, after a pixel three primary color components are extracted, the RGB that will obtain
Component is further converted into HSV (tone Hue, saturation degree Saturation, brightness Value) value, and compares
It is to judge that the color of the pixel is predetermined color compared with its tone value whether within the scope of default tone value,
Otherwise the color for judging the pixel is non-predetermined color.
Then computing unit 14 calculates the pixel quantity with predetermined color and accounts for pixel total in a cell
The ratio (hereinafter referred to as accounting) of quantity.Comparing unit 15 again accounting compared with one first preset value, if
The accounting exceedes or reached the first preset value, and it is predetermined color to assert that the color of the cell is shown.First is pre-
If value can be 50%, 60% or other numerical value.Other first preset value can be stored in memory cell 16
In.
It can identify that the cell belongs to working region or inoperative in conjunction with some other information of cell
Region.The texture information of cell is referred in the present embodiment.Main control module 3 also includes texture blending list
Member 17, texture comparing unit 18.Texture blending unit 17 extracts the texture eigenvalue of the cell.Unit
The dispersion of at least one parameter of all pixels can embody the difference between each value of the parameter in lattice
Degree.If target area is the paint of green, the dispersion very little of a parameter in its image, even
0.Due to the crossgrained on meadow, the dispersion meeting of the difference value of a parameter of all pixels of cell
More than or equal to one default dispersion, so as to embody the scrambling of the texture of the cell.Therefore,
In present embodiment, the texture eigenvalue is parameter discrete degree, as color dispersion, gray scale dispersion,
Brightness dispersion etc..
Texture comparing unit 18 compares the texture eigenvalue of the cell with the second preset value to judge texture spy
Whether value indicative reaches the second preset value.In present embodiment, second preset value is default dispersion.Line
The presence that comparing unit 18 can be independent is managed, comparing unit 15 can also be integrated into.And the second preset value
It can be stored in advance in memory cell 16.
Main control module 3 also includes recognition unit 19.Wherein color extraction unit 13, computing unit 14, ratio
Compared with the part that unit 15 and memory cell 16 may be constructed recognition unit 19 in one embodiment, or
Person says that being integrated into recognition unit 19 forms entirety.Can also in another embodiment as with recognition unit
19 unit blocks arranged side by side.The pixel with predetermined color in the cell is identified when recognition unit 19
When the texture eigenvalue that accounting meets or exceeds the first preset value and cell meets or exceeds the second preset value,
Judge that travel region is in target area corresponding to the cell；If accounting is not up to the first preset value or texture is special
Value indicative is not up to the second preset value, judges target area corresponding to the cell for non-travel region.Therefore control
The method of automatic running device processed further comprises step S30, that is, read the identification letter that each cell is included
Cease simultaneously it is identified, so as to target area corresponding to obtaining unit lattice whether be working region identification knot
Fruit.
All cells in image are all identified the recognition unit 19 of main control module 3 respectively, so as to obtain
Obtain the recognition result of all cells.In embodiment preferably, main control module 3 also includes amending unit
32, amending unit 32 is modified based on markov stochastic model to the recognition result of cell.Therefore exist
In the embodiment, therefore control method further comprises step S60, i.e., the mode amendment list based on smoothing processing
Different recognition result in first lattice.Because in actual condition, the recognition result that is obtained by step S30
Certain error is had, that is, produces abnormal recognition result.Makeover process can be the abnormal identification
As a result it is modified, so as to lift the accuracy of identification.Specifically, for each unit in image
For lattice, cell adjacent thereto is certainly existed.Utilize the recognition result of its adjacent cell and right
The recognition result of cell itself carries out considering the purpose that can reach amendment.Amending unit 32 includes
Information extraction unit 20 and information change unit 21.
In one embodiment, the method for amendment is as follows：Step S60 includes step S61, S62, S63
And S64.Wherein step S61 refers to for each cell, information extraction unit 20 extract it is all with
The recognition result of the adjacent cell of the cell；Step S62 refers to the statistics of computing unit 14 and the list
The quantity of the recognition result identical adjacent cells lattice of first lattice, and the quantity account for adjacent cell sum
The accounting of amount.For example, if the recognition result of the cell is working region, computing unit 14 counts
Recognition result is also the quantity of working region in cell adjacent thereto, and calculates the quantity and account for entirely
The accounting of adjacent cell total quantity.Step S63 refers to that comparing unit 15 compares the accounting and the 4th pre-
If the size of value, if the accounting is more than or equal to one the 4th preset value, (the 4th preset value is many under normal circumstances
Can be 50%, 75% etc. in 50%), illustrate that the adjacent cells lattice of the recognition result occupy all adjacent lists
Most of first lattice, therefore changing unit 21 by cell information makes the recognition result of the cell keep not
Become.If the accounting is less than the 4th preset value, cell information, which changes unit 21, makes the identification knot of the cell
Fruit is changed to another recognition result, such as the script recognition result of the cell is working region, after change i.e.
Recognition result is changed into non-active area.Whole process for example, for a cell, originally
Recognition result be non-active area.But there is the identification of 3 cells in 4 cells adjacent thereto
As a result it is working region, the ratio (3/4=75%) has been above the 4th preset value (assuming that the 4th preset value is
50%), then the original recognition result of the cell and neighbouring unit are assert according to the result of the adjacent cells lattice
The recognition result of lattice is identical, therefore the recognition result of the cell keeps constant, remains as working region.If
The recognition result for there was only 1 cell in 4 cells adjacent thereto is working region, the ratio
(1/4=25%) is less than the 4th preset value, then assert the identification of the recognition result cell adjacent thereto of the cell
As a result do not meet, may be to caused by the recognition result of the cell be due to error, therefore the cell
Recognition result be modified to working region.Certainly, finally it is to be emphasized again that herein adjacent is not to limit up and down
4 directions it is adjacent, totally 8 directions such as left, upper right, bottom left, bottom right can also be not limited to.Together
The original recognition result of sample cell is not limited to working region or non-active area.Final step S64
Refer to be adapted in this approach on all cells, that is, complete the modified result of whole image, that is to say, that
Above-mentioned S61~S63 steps are performed to all cells, correct the recognition result of all cells.
In other one embodiment, step S60 include step S66, S67, S68, S69, S70,
S71 and S72.The reliability of the recognition result of step S66 elder generations acquiring unit lattice.Reliability is usually 0~100%
Between a numerical value.Certainly, reliability can also be the numerical value of other forms.Reliable in step S67
Degree is designated as Y1, and unreliable degree is designated as N1, wherein N1=1-Y1.Y1 can also be referred to as similar reliability.
N1 can be referred to as dissimilar reliability.Y1 and N1 can be stored in memory cell 16.Then again
The reliability for calculating the cell adjacent with the cell is obtained by step S68.Reliability is usually 0~100%
Between a numerical value.If if the cell has 8 cells adjacent thereto, pass through step S69
Acquisition 8 similar reliabilitys similar with step S67 and the reliability of dissmilarity.Specifically, this
The reliability of 8 adjacent cells is designated as similar reliability Ya, Yb, Yc ..., and dissmilarity
Reliability is designated as Na, Nb, Nc ....Then this 8 similar reliabilitys are added by step S70
Power summation process obtains Y2.In the present embodiment, weight coefficient is in the same size, is preferably 1/8.When
So, weight coefficient can also use mutually different value.Similarly, that 8 dissimilar reliabilitys are carried out
Weighted sum handles to obtain N2.The weight coefficient of 8 dissimilar reliabilitys can be the same, and
Can be consistent with the weight coefficient of similar reliability.Then Y1+ α N1 and Y2+ are compared by step S71
α N2 size simultaneously takes corresponding measure.Specifically, α is weight coefficient here, and and above several steps
Rapid weighted value can be with different, can also be the same.Further, the process compared can be in comparing unit
Carry out, can also be carried out in other elements in 15.On result of the comparison, if Y1+ α N1 result
More than or equal to Y2+ α N2 result, then information changes unit 21 and maintains the recognition result of the cell not
Become, if Y1+ α N1 result is less than Y2+ α N2 result, information changes unit 21 and changes the unit
The recognition result of lattice.Then said process is all performed to all cells in image by step S72,
Each cell can participate in iterative cycles, until the recognition result of all cells no longer produces change.
Therefore the method for control automatic running device further comprises step S40, and step S40 includes for judgement
Whether target area corresponding to the subimage block of several cells is travel region.Automatic running device 1
Including for performing the judging unit 22 of the step.Judging unit 22 includes subimage block division unit 23,
Subimage block division unit 23 is used to divide the image into several subimage blocks.In one of embodiment,
The mode specifically divided is as follows：Step S40 includes step S41, S42, S43.First, step is passed through
If S41 selectively divides an image into according to the direction of travel subimage block division unit 23 of automatic running device
Dry subimage block.Each subimage block corresponds to different direction of travel.In one of which embodiment,
Subimage block division unit 23 divides the image into three middle part, left part and right part subimage blocks, corresponds to respectively
Subregion in target area.As shown in figure 3, middle part correspond to automatic running device 1 front dead center,
The intermediate region a wide with automatic running device 1；Left part corresponds to front, the position of automatic running device 1
Left field b on the left of a of intermediate region；Right part corresponds to the front of automatic running device 1, is located at
Between right side area c on the right side of a of region.Three subimage blocks each contain multiple cells respectively.
In other one embodiment, subimage block division unit 23 can also divide the image into front, it is left front
5 different subimage blocks such as side, left, right front, right.Because each subimage block contains
Some cells, judging unit 22 judge the son by the recognition result of all cells in subimage block
Target area corresponding to image block is that travel region is also non-travel region.Specifically, it is assumed that with positioned at image
Totally 60 cells are used as middle part subimage block to three row cells of front end.In this embodiment, automatically walk
The information extraction unit 20 of equipment 1 extracts the recognition result of all cells in the middle part subimage block, calculates
Unit 14 calculates quantity of the recognition result for the cell in working region region, and is A the number tag.
Certainly the quantity for the cell that recognition result is non-active area can also be counted in other embodiments.Compare
Unit 15 compares the quantity and the size of the 3rd preset value for the cell that recognition result is travel region.Work as quantity
When the ratio that A or A accounts for all cells in the subimage block is more than or equal to three preset value, then
Judging unit 22 can assert that the subimage block is travel region.It can certainly be arranged to when recognition result is non-
When the quantity of the cell of travel region is less than three preset value, can determine whether the subimage block is travel region.
The 3rd preset value in the present embodiment is pre-stored in memory cell 16, and can be the numbers such as 30,40,50
Value.In other embodiments, automatic running device 1 can also be travel region or non-walking recognition result
The cell in region accounts for the ratio of all cells of the subimage block as parameter, default with the another 3rd
Value is compared, and the 3rd preset value in the present embodiment is more than or equal to 50%, can be 50%, 60%,
90% etc..
After target area corresponding to subimage block is judged for travel region or non-travel region, automatically walk is set
Standby 1, by step S50, controls automatic running device to advance, retreats, turns left or turn right according to judged result.
According to the judged result of judging unit 22, automatic running device 1 can perform specific response action.Walking mould
Block 4 control automatic running device 1 respond action have including：Advance (F), retreat (B), turn left (L), be right
Turn (R) and do not change (N).Image is being divided into the embodiment of the three pieces of subimage blocks in left, center, right, due to every
The recognition result of individual subimage block has travel region and non-travel region respectively.So eight kinds of different feelings altogether
Condition：It is that 1. left, center, rights are travel region respectively；2. left is travel region, the right side is non-travel region；3.
Left and right is travel region, in be non-travel region；4. a left side is travel region, the middle right side is non-travel region；5.
A left side is non-travel region, and the middle right side is travel region；6. or so be non-travel region, in be travel region；7.
Left is non-travel region, and the right side is travel region；8. left, center, right is non-travel region.
In the case of the 1st kind, main control module 3 makes walking module 4 perform the action for not changing (N)；
In case 2, main control module 3 makes walking module 4 perform and the action of (LF) of advancing；
In the case of the 3rd kind, main control module 3 makes walking module 4 perform the dynamic of (BLF) that retreat and advance
Make；
In the case of the 4th kind, main control module 3 makes walking module 4 perform the dynamic of (BLF) that retreat and advance
Make；
In the case of the 5th kind, main control module 3 execution of walking module 4 is turned right and the action of (RF) of advancing；
In the case of the 6th kind, main control module 3 makes walking module 4 perform retrogressing right-hand rotation and advances (BRF)
Action；
In the case of the 7th kind, main control module 3 makes walking module 4 perform retrogressing right-hand rotation and advances (BRF)
Action；
In the case of the 8th kind, main control module 3 make walking module 4 perform retreat turn right and advance (BRF) or
Person retreats and the action for (BLF) alternative that advances.
Further illustrate, when it is travel region to judge current goal region, automatic running device 1 can
It can may proceed to perform original Running strategy, such as keep original walking states；When identification current goal area
When domain is non-travel region, automatic running device 1 can then change direction of travel, further can be selectively
Walked to the direction away from the subimage block.Because multiple subimage blocks, automatic running device 1 be present in image
Need to carry out the plurality of subimage block respectively the identification of travel region or non-travel region, and then take correspondingly
Strategy.In a preferred embodiment, automatic running device can be recognized the plurality of subimage block simultaneously
It is fixed.For example, for three subimage blocks such as middle part, left part, right part, if detecting this three
Subimage block is travel region, the state that automatic running device maintenance moves forward；If detect this
Three subimage blocks are non-travel region, and automatic running device can turn to 180 degree, rearward move；If
Detect that middle part, the subimage block of left part are non-travel region, right part is travel region, then automatically walk
Equipment will move to the right to the direction away from middle part and left part, it is of course possible to first retreat turn right again or
Person first turns right a variety of specific modes such as retreats again.
Certainly in a preferred embodiment, the subimage block division unit 23 of automatic running device 1 can also wrap
Repeatedly division subimage block is included, then carries out the process of comprehensive descision.Corresponding to the subimage block divided every time
Region can differ.So the judged result of different zones is considered, avoids single area
The inaccuracy of judged result and the mistake of policy development brought, improve the walking of automatic running device 1
Accuracy.Specifically still positioned at three row cells of graphics front-end, totally 60 cells are used as middle part subgraph for act
As exemplified by block, once judging that the object of the subimage block is 60 cells, what is once identified in addition
During, can using four row cells of graphics front-end, totally 80 cells are as middle part subimage block, at this
The secondary object for judging the subimage block is 80 cells.And the 3rd preset value used when judging twice
Also differ, herein can with but be not limited to 60.This is identified twice and combined as brand-new judgement
Foundation, such as when Rule of judgment has 40 cells to be identified as walking to form in 60 cells of three rows
Region, and form four rows 80 cells in there are 60 cells to be identified as travel region when, can recognize
Fixed middle part is travel region.If two Rule of judgment can not meet simultaneously, assert that middle part is non-walking area
Domain.Certainly, can also judge in the same way for left part and right part.
In an additional preferred embodiment, because individual image still is possible to adopt the information of target area
Collection, which has, has distortion, for example, can be produced in certain object quickly skimmed in a flash to target area cloudy
Shadow, so as to influence deterministic process of the automatic running device to the target area.Therefore in the present embodiment, certainly
Dynamic walking arrangement is additionally included in setting steps S80 between above-mentioned steps S40 and step S50.Pass through step S80
Subimage block in multiple images carries out synthetic filter, obtains whether the subimage block is travel region
Final judged result.Can be more so as to be formed within cycle regular hour to the multiple shooting in target area
The image of frame.Then the judgement information included to each two field picture carries out synthetic filter and obtains final judgement knot
Fruit.
Method and step S80 has comprised at least step S81, S82 and step S84.In a preferred embodiment,
Step S80 further comprises the step S83 between S82 and S84.Specific method is as follows:Pass through step
S81, to same target area, repeatedly shooting forms multiple image, each two field picture to image collecting device 2
It is referred to as the first two field picture, the second two field picture ... nth frame image.Automatic running device 1 also includes note
Record unit 33, recording unit 33 is used for according to being carried out to the judged result of subimage block pair by step S82
Weighted value is handled.Specifically, when it is travel region that judging unit 22, which judges the first two field picture,
Recording unit 33 can add one the 5th preset value to initial weighted value.Initial weighted value for convenience of explanation,
It can mark and be, naturally it is also possible to labeled as other numerical value.And the 5th preset value can be default fixation
Constant, and the function of change.In the present embodiment, the 5th preset value can be but be not limited to 3.When
During the recognition result travel region of the first two field picture, recording unit just makes corresponding weighted value just be changed into 3.So
Handled afterwards in the recognition result to the second two field picture.If the recognition result of the second two field picture is also walking area
During domain, the recording unit 33 of automatic running device 1 adds one the 5th preset value to current weighted value.This
When corresponding weighted value be just changed into 6.If the recognition result of the second two field picture is not travel region, record
Unit 33 is not done to present weight value to be changed.Then the recognition result of the 3rd two field picture is handled again.If
When the recognition result of 3rd two field picture is also travel region, then present weight value is changed into 9.So it is continued until
Nth frame image.It is separately also default to current weighted value and one the 7th by step S84, comparing unit 15 again
Value is compared.When current weighted value is more than or equal to seven preset values, it is assumed that judged result is correct
, i.e., certain current goal region is travel region.Wherein the 7th preset value can for example be set to 8.Pass through
This mode considers to the recognition result of multiple image, so as to avoid the mistake that may be present of single-frame images
The harmful effect that result is brought by mistake.Further, the identification to each two field picture can be resolved into image
The identification of each subimage block.For example, each two field picture can resolve into the subimage block of left, center, right three, note
Record unit can record to three subimage blocks respectively, correspond to three weighted values respectively.
In addition in a preferred embodiment, step S83 is additionally included in the image handoff procedure of frame and frame, note
Record unit 33 can also subtract one the 6th preset value to current weighted value, so that present weight value reaches or surpassed
Crossing the 7th preset value, to obtain the time elongated so that the image of more frame numbers is considered, and is further improved
Accuracy.In the present embodiment, the 6th preset value can with but be not limited to be 1.Such as first two field picture identification
As a result it is travel region, then present weight value is changed into 3.When the second two field picture recognition result is non-travel region
When, present weight value is changed into 2.When the 3rd two field picture recognition result is travel region, present weight value is just
It is changed into 4.When the 4th two field picture recognition result is travel region, present weight value is just changed into 6.So until
When weighted value meets or exceeds seven preset values, assert that current goal region is travel region.If weighted value
When being not up to seven preset values all the time, assert that current goal region is non-travel region.Certainly, this area skill
Art personnel it is conceivable that by some change, when can make it that weighted value meets or exceeds seven preset values,
Assert that current goal region be non-travel region, and when weighted value is not up to seven preset values all the time, assert ought
Preceding target area is travel region.The identification condition of travel region and non-travel region is exchanged.And
And the computation rule of weighted value is further refined, if such as can set when weighted value subtracts in any condition
To minimum value, such as when weighted value is reduced to 0, would not continue to reduce again.
The present invention is not limited to lifted specific embodiment structure, and the structure based on present inventive concept belongs to this
Invention protection domain.
Claims (18)
- A kind of 1. method for controlling automatic running device walking, it is characterised in that：It the described method comprises the following steps：S10, the image for obtaining the automatic running device steps target region；S20, several cells are divided the image into, each cell has at least one adjacent cells lattice；S30, the colouring information of specified pixel and the textural characteristics of the cell in the cell Value identifies whether target area corresponding to the cell is working region, and obtains recognition result；S40, several subimage blocks are divided the image into, each subimage block includes several adjacent units Lattice, the mesh according to corresponding to the recognition result of the cell in the subimage block judges the subimage block Mark whether region is walkable region, and obtain judged result；S50, according to judged result, control the direction of travel of automatic running device.
- 2. according to the method for claim 1, it is characterised in that：Methods described is additionally included in acquisition identification knot After fruit, for each cell, according to the cell and its recognition result of adjacent cells lattice, adjust The recognition result of the whole cell.
- 3. according to the method for claim 2, it is characterised in that：Methods described further comprises the steps：S61, a cell is selected, and obtain its recognition result；S62, statistics are with the quantity with the adjacent cells lattice of same identification result in step S61；Ratio of the quantity in total adjacent cells lattice quantity in S63, calculation procedure S62；If S64, the ratio exceed or reached preset value, the identification knot for the cell that maintenance step S61 is specified Fruit is constant；If the ratio is less than preset value, change the recognition result for the cell that step S61 is specified, Wherein described 4th preset value is more than or equal to 50%.
- 4. according to the method for claim 3, it is characterised in that：The adjacent cells lattice include horizontal and vertical The cell adjacent with the selected cell upwards.
- 5. according to the method for claim 4, it is characterised in that：The adjacent cells lattice also include with it is described Transverse direction and longitudinal direction cell adjacent with the selected cell on the direction in 45 degree of angles.
- 6. according to the method for claim 2, it is characterised in that：Methods described further comprises the steps：S66, a cell is selected, obtain reliability Y1 of the cell for its recognition result, can A numerical value between being 0~100% by degree Y1；S67,1-Y1 is calculated, and be N1 by result queue；S68, reliability Ya of all adjacent cells lattice for its recognition result for obtaining the selected cell, Yb ..., reliability Ya, Yb ... are a numerical value between 0~100%；S69,1-Ya, 1-Yb ... are calculated, and be Na, Nb ... result queue；S70, Ya, Yb ..., weighted sum obtains weighted sum Y2, and the weighted sum Nb, Nc ... is added Power and be N2, wherein, weight coefficient all same；S71, Y1+ α N1 and Y2+ α N2 result being calculated respectively and compares its size, wherein α is coefficient；If S72, Y1+ α N1 result are more than or equal to Y2+ α N2 result, remain described and specify The recognition result of cell is constant, if Y1+ α N1 result is less than Y2+ α N2 result, changes The recognition result of the cell specified.
- 7. according to the method for claim 1, it is characterised in that：The S40 steps further comprise following Step：S41, described image is divided into several subimage blocks, obtains the cell that each subimage block includes Quantity, be marked as B；S42, the recognition result for collecting cell in the subimage block, statistics recognition result is working region Cell quantity, and be marked as A；If S43, A：B is less than the 3rd preset value, then judges that target area corresponding to the subimage block is not Walkable region, otherwise, judge that walkable region is in target area corresponding to the subimage block.
- 8. according to the method for claim 1, it is characterised in that：Methods described also includes to same target area Domain is continuously shot to form multiple image, is judged according to the judged result of same subimage block in every two field picture Whether target area corresponding to the subimage block is walkable region, and obtains judged result.
- 9. according to the method for claim 8, it is characterised in that：Methods described further comprises the steps：S81, same target area is continuously shot to form multiple image；A subimage block in S82, a selected wherein two field picture, judged result is obtained by step S40；S83, initial parameter value is set, and parameter value transported according to the step S82 judged results obtained Calculate, if judged result is walkable region, increase and judged result pass in the initial parameter value First parameter of connection turns into current parameter value；If judged result is not walkable region, described in holding Parameter value is constant；S84, next two field picture is selected, and current parameter value is entered according to the step S82 judged results obtained Row computing, if judged result is walkable region, increase and judgement on the current parameter value As a result the first parameter associated turns into new current parameter value；If judged result is not walkable region, Then keep the current parameter value constant；The size of the current parameter value of S85, comparison and threshold value, if current parameter value is more than or equal to threshold value, Then assert that target area corresponding to the subimage block is walkable region.
- 10. according to the method for claim 9, it is characterised in that：The step S84 is still further comprised： After selected next two field picture before computing is carried out to current parameter value, current parameter value subtracts one Default second parameter, and second parameter is less than first parameter.
- 11. according to the method for claim 1, it is characterised in that：The subimage block includes middle part, left part With three subimage blocks of right part, intermediate region, left field and the right side region of target area are corresponded to respectively Domain.
- A kind of 12. automatic running device, it is characterised in that：Image collecting device including housing, on housing, Described image harvester is used for photographic subjects region and generates image, drives the automatic running device The walking module of walking, described image harvester and walking module are connected to control automatic running device The main control module of work, wherein, the main control module includes division unit, recognition unit, judges list Described image is divided into several cells by member and control unit, the division unit, and division As a result the recognition unit is passed to, the recognition unit identifies target area corresponding to the cell Whether it is working region, and recognition result is passed to judging unit, judging unit is judged comprising some Whether the subimage block corresponding region of cell is walkable region, and judged result is passed to control Unit, described control unit control the direction of travel of walking module according to judged result.
- 13. automatic running device according to claim 12, it is characterised in that：The main control module also includes Amending unit, the amending unit are directed to each cell, according to cell and its adjacent cells lattice Recognition result, adjust the recognition result of the cell.
- 14. automatic running device according to claim 13, it is characterised in that：The adjacent cells lattice include The cell adjacent with the selected cell in transverse direction and longitudinal direction.
- 15. automatic running device according to claim 14, it is characterised in that：The adjacent cells lattice also wrap Include and unit adjacent with the selected cell on direction of the transverse direction and longitudinal direction in 45 degree of angles Lattice.
- 16. automatic running device according to claim 12, it is characterised in that：The judging unit also includes Subimage block division unit, the subimage block division unit divide the image into several subimage blocks, The recognition result for the cell that judging unit includes according to the subimage block judges that corresponding subimage block is No is walkable region.
- 17. automatic running device according to claim 16, it is characterised in that：During the subimage block includes Three portion, left part and right part subimage blocks.
- 18. automatic running device according to claim 12, it is characterised in that：The main control module also includes Record has the recording unit of initial parameter value, and described image harvester is continuously clapped same target area Take the photograph to form multiple image, the judging unit judges the same subimage block in each two field picture And judged result is obtained, recording unit carries out computing according to its judged result to parameter value, works as parameter value More than or equal to threshold value, then assert that target area corresponding to the subimage block is walkable region.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610389387.3A CN107463166A (en) | 2016-06-03 | 2016-06-03 | Automatic running device and its control traveling method |
| PCT/CN2017/087021 WO2017206950A1 (en) | 2016-06-03 | 2017-06-02 | Automatic walking device and method for controlling the walking thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610389387.3A CN107463166A (en) | 2016-06-03 | 2016-06-03 | Automatic running device and its control traveling method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107463166A true CN107463166A (en) | 2017-12-12 |
Family
ID=60545561
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610389387.3A Pending CN107463166A (en) | 2016-06-03 | 2016-06-03 | Automatic running device and its control traveling method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107463166A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019158090A1 (en) * | 2018-02-13 | 2019-08-22 | 苏州宝时得电动工具有限公司 | Automatic work system, and self-mobile device and control method thereof |
| CN110612806A (en) * | 2018-06-19 | 2019-12-27 | 灵动科技（北京）有限公司 | Intelligent mower |
| CN111339961A (en) * | 2020-02-28 | 2020-06-26 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| CN111353431A (en) * | 2020-02-28 | 2020-06-30 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| CN112558595A (en) * | 2019-09-06 | 2021-03-26 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| CN112884850A (en) * | 2021-03-22 | 2021-06-01 | 上海易教科技股份有限公司 | Method and device for acquiring picture for track identification |
| CN113495552A (en) * | 2020-03-19 | 2021-10-12 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| WO2021243895A1 (en) * | 2020-06-02 | 2021-12-09 | 苏州科瓴精密机械科技有限公司 | Image-based working position identification method and system, robot, and storage medium |
| CN113805571A (en) * | 2020-05-29 | 2021-12-17 | 苏州科瓴精密机械科技有限公司 | Robot walking control method and system, robot and readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101331515A (en) * | 2005-12-21 | 2008-12-24 | 日本电气株式会社 | Gray-scale correcting method, gray-scale correcting device, gray-scale correcting program, and image device |
| CN101739551A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Method and system for identifying moving objects |
| CN103020222A (en) * | 2012-12-13 | 2013-04-03 | 广州市香港科大***研究院 | Visual mining method for vehicle GPS (global positioning system) data analysis and abnormality monitoring |
| CN103903238A (en) * | 2014-03-21 | 2014-07-02 | 西安理工大学 | Method for fusing significant structure and relevant structure of characteristics of image |
| CN104111653A (en) * | 2013-04-22 | 2014-10-22 | 苏州宝时得电动工具有限公司 | Automatic walking equipment and working region judgment method thereof |
| CN105512689A (en) * | 2014-09-23 | 2016-04-20 | 苏州宝时得电动工具有限公司 | Lawn identification method based on images, and lawn maintenance robot |
-
2016
- 2016-06-03 CN CN201610389387.3A patent/CN107463166A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101331515A (en) * | 2005-12-21 | 2008-12-24 | 日本电气株式会社 | Gray-scale correcting method, gray-scale correcting device, gray-scale correcting program, and image device |
| CN101739551A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Method and system for identifying moving objects |
| CN103020222A (en) * | 2012-12-13 | 2013-04-03 | 广州市香港科大***研究院 | Visual mining method for vehicle GPS (global positioning system) data analysis and abnormality monitoring |
| CN104111653A (en) * | 2013-04-22 | 2014-10-22 | 苏州宝时得电动工具有限公司 | Automatic walking equipment and working region judgment method thereof |
| CN103903238A (en) * | 2014-03-21 | 2014-07-02 | 西安理工大学 | Method for fusing significant structure and relevant structure of characteristics of image |
| CN105512689A (en) * | 2014-09-23 | 2016-04-20 | 苏州宝时得电动工具有限公司 | Lawn identification method based on images, and lawn maintenance robot |
Non-Patent Citations (2)
| Title |
|---|
| 宓超 等: "《装卸机器视觉及其应用》", 31 January 2016, 上海科学技术出版社 * |
| 彭真明 等: "《光电图像处理及应用》", 30 April 2013, 电子科技大学出版社 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019158090A1 (en) * | 2018-02-13 | 2019-08-22 | 苏州宝时得电动工具有限公司 | Automatic work system, and self-mobile device and control method thereof |
| CN110612806A (en) * | 2018-06-19 | 2019-12-27 | 灵动科技（北京）有限公司 | Intelligent mower |
| US11327502B2 (en) | 2018-06-19 | 2022-05-10 | Lingdong Technology (Beijing) Co. Ltd | Smart lawn mower |
| CN110612806B (en) * | 2018-06-19 | 2021-04-20 | 灵动科技（北京）有限公司 | Intelligent mower |
| CN112558595A (en) * | 2019-09-06 | 2021-03-26 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| CN111353431A (en) * | 2020-02-28 | 2020-06-30 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| WO2021169193A1 (en) * | 2020-02-28 | 2021-09-02 | 苏州科瓴精密机械科技有限公司 | Automatic working system, automatic locomotion device and control method therefor, and computer readable storage medium |
| WO2021169192A1 (en) * | 2020-02-28 | 2021-09-02 | 苏州科瓴精密机械科技有限公司 | Automatic working system, automatic walking device and control method therefor, and computer-readable storage medium |
| CN111339961A (en) * | 2020-02-28 | 2020-06-26 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| CN111353431B (en) * | 2020-02-28 | 2024-03-08 | 苏州科瓴精密机械科技有限公司 | Automatic working system, automatic walking equipment, control method thereof and computer readable storage medium |
| CN113495552A (en) * | 2020-03-19 | 2021-10-12 | 苏州科瓴精密机械科技有限公司 | Automatic work system, automatic walking device, control method thereof, and computer-readable storage medium |
| CN113805571A (en) * | 2020-05-29 | 2021-12-17 | 苏州科瓴精密机械科技有限公司 | Robot walking control method and system, robot and readable storage medium |
| CN113805571B (en) * | 2020-05-29 | 2024-03-12 | 苏州科瓴精密机械科技有限公司 | Robot walking control method, system, robot and readable storage medium |
| WO2021243895A1 (en) * | 2020-06-02 | 2021-12-09 | 苏州科瓴精密机械科技有限公司 | Image-based working position identification method and system, robot, and storage medium |
| CN112884850A (en) * | 2021-03-22 | 2021-06-01 | 上海易教科技股份有限公司 | Method and device for acquiring picture for track identification |
| CN112884850B (en) * | 2021-03-22 | 2024-02-13 | 上海易教科技股份有限公司 | Method and device for acquiring pictures for track identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107463166A (en) | Automatic running device and its control traveling method | |
| CN107463167A (en) | Automatic running device and target area recognition methods | |
| CN205692049U (en) | A kind of grass-removing robot in boundless boundary line | |
| CN104111653A (en) | Automatic walking equipment and working region judgment method thereof | |
| CN104239886B (en) | The extracting method on lawn and background line of demarcation based on graphical analysis | |
| CN105785986A (en) | Automatic working equipment | |
| Åstrand et al. | An agricultural mobile robot with vision-based perception for mechanical weed control | |
| CN106155053A (en) | A kind of mowing method, device and system | |
| WO2021169193A1 (en) | Automatic working system, automatic locomotion device and control method therefor, and computer readable storage medium | |
| CN110297483A (en) | To operating area boundary acquisition methods, device, operation flight course planning method | |
| CN110243372B (en) | Intelligent agricultural machinery navigation system and method based on machine vision | |
| CN104345734A (en) | Automatic working system, automatic walking equipment and control method thereof | |
| EP2922384B1 (en) | Autonomous locomotion device | |
| CN107239074A (en) | Automatic working system and its map method for building up of working region | |
| WO2022021630A1 (en) | Autonomous walking device and control method and system therefor, and readable storage medium | |
| CN105069431B (en) | The localization method and device of face | |
| WO2021169192A1 (en) | Automatic working system, automatic walking device and control method therefor, and computer-readable storage medium | |
| CN107564071A (en) | A kind of image recognition meadow method and device | |
| EP3413155B1 (en) | Method for the detection of at least one section of a limiting edge of a surface to be processed, method for operating an autonomous mobile green area processing robot, detection system and green area processing system | |
| CN111460903A (en) | System and method for monitoring growth of field broccoli based on deep learning | |
| CN102208099A (en) | Illumination-change-resistant crop color image segmentation method | |
| CN109387214A (en) | A kind of Robot Path Planning Algorithm based on virtual wall | |
| CN112020982A (en) | Automatic lodging crop harvesting method and harvester | |
| Mangaonkar et al. | Fruit harvesting robot using computer vision | |
| WO2021042487A1 (en) | Automatic working system, automatic travelling device and control method therefor, and computer readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |