Summary of the invention
The object of the present invention is to provide a kind of dispatch control method of transportation robot and system, realize that whole process does not need to adjust
Degree side participates in that transportation dispatching can be completed, and reduces the retardance of transport task caused by participating in because of dispatching party, and then promotion pair
Efficiency and cargo transportation efficiency are connect, application popularization rate is improved.
Technical solution provided by the invention is as follows:
The present invention provides a kind of dispatch control method of transportation robot, comprising steps of
It detects itself working condition and obtains bulk state information;
Shooting obtains target image, and the target image includes transmission device, and is set at the transmission device, and uses
In the indicator light for showing transmission device working condition;
The target image is analyzed to obtain the work state information of the transmission device;
According to the bulk state information and the work state information, analysis obtains the target docked with itself transmission dress
Set and convey task type;
The spatial position of the target transmission device is obtained, and navigates and is moved to target transmission device position
Place;
It is docked according to the conveying task type with the target transmission device, completes the loading or unloading of cargo.
Further, the work state information for being analyzed to obtain the transmission device to the target image is specific
Comprising steps of
By visual detection algorithm, LED status knowledge is carried out to the target image using default neural network model
Not;
According to LED status recognition result, analysis obtains the work state information of the transmission device;
Wherein, the LED status includes instruction lamp on/off state and indicator light shape and color state;The indicator light shape
Color state includes indicator light colors state and indicator light shape state.
Further, the target image is analyzed after obtaining the work state information of the transmission device, root
According to the bulk state information and the work state information, analysis obtains target transmission device and the conveying docked with itself
Before task type comprising steps of
The shared work state information for respectively analyzing obtained transmission device of transmission is carried out with remaining transportation robot.
Further, described according to the bulk state information and the work state information, it is right with itself that analysis is obtained
The target transmission device and conveying task type connect specifically includes step:
Judge that itself is in loading condition or state to be unloaded according to the bulk state information;
When itself be in when loading condition, determine the work state information be the corresponding transmission device of state to be sent
For target transmission device, and determine that the conveying task type is cargo reception type;
When itself is in state to be unloaded, determine that the work state information is the corresponding transmission device of state to be received
For target transmission device, and determine that the conveying task type is cargo dispensing type.
Further, it further comprises the steps of:
When according to the bulk state information and the work state information, analysis obtains at least two work with itself
When the candidate transmission device that state matches, the distance between itself and each candidate transmission device value are calculated;
The size for comparing all distance values determines that the minimum corresponding candidate transmission device of distance value is target transmission dress
It sets.
Further, the spatial position for obtaining the target transmission device, and navigate and be moved to the target transmission
Device position specifically includes step:
At least four target languages of target transmission device in the target image are identified by visual detection algorithm detection
Adopted point;The target semanteme point is immobilizes on the target transmission device, and the point that identification is high;
According to the dimension information of the target transmission device, the first space bit of the target transmission device is calculated
It sets;
The target transmission device position is moved to according to first spatial position navigation.
Further, the spatial position for obtaining the target transmission device, and navigate and be moved to the target transmission
Device position specifically includes step:
At transmitting detection laser to the stabilizer blade of the target transmission device, obtains each stabilizer blade and swash what laser coordinate was fastened
Light coordinate;
The second space position of the target transmission device is calculated according to the laser coordinate;
The target transmission device position is moved to according to second space position navigation.
The present invention also provides a kind of Dispatching Control Systems of transportation robot, comprising: several transportation robots and transmission
Device;The transmission device is equipped with the indicator light for showing itself working condition;Each transportation robot includes: image
Acquisition module, processing module, detection module, analysis module, control module and execution module;
The detection module obtains bulk state information for detecting itself working condition;
Described image acquisition module obtains target image for shooting, and the target image includes transmission device, Yi Jishe
Indicator light at the transmission device;
The processing module is connect with described image acquisition module, obtains institute for being analyzed the target image
State the work state information of transmission device;
The analysis module is connect with the processing module and the detection module respectively, for according to the ontology shape
State information and the work state information, analysis obtain the target transmission device docked with itself and conveying task type;
The control module is connect with the analysis module, for obtaining the spatial position of the target transmission device, and
Navigation is moved to the target transmission device position;
The execution module is connect with the analysis module, for being moved to the target transmission device position
Afterwards, it is carried out docking the loading or unloading for completing cargo with the target transmission device according to the conveying task type.
Further, the processing module includes: the first image identification unit and first processing units;
The first image recognition unit, for passing through visual detection algorithm, using default neural network model to described
Target image carries out LED status identification;
The first processing units are connect with the first image recognition unit, are tied for being identified according to LED status
Fruit, analysis obtain the work state information of the transmission device;
Wherein, the LED status includes instruction lamp on/off state and indicator light shape and color state;The indicator light shape
Color state includes indicator light colors state and indicator light shape state.
Further, each transportation robot further include: wireless communication module;
The wireless communication module is connect with the wireless communication module of remaining transportation robot, shared for carrying out transmission
Respectively analyze the work state information of obtained transmission device.
Further, the analysis module includes: judging unit and the first determination unit;
The judging unit, for judging that itself is in loading condition or to be unloaded according to the bulk state information
State;
First determination unit is connect with the judging unit, for determining the work when itself being in when loading condition
Be the corresponding transmission device of state to be sent as status information it is target transmission device, and determines that the conveying task type is goods
Object receives type;When itself is in state to be unloaded, determine that the work state information is the corresponding transmission of state to be received
Device is target transmission device, and determines that the conveying task type is that cargo launches type.
Further, the analysis module further include: the second processing unit and the second determination unit;
Described the second processing unit is analyzed for working as according to the bulk state information and the work state information
To at least two match with itself working condition candidate transmission device when, calculate itself and each candidate transmission device it
Between distance value, the size of more all distance values;
Second determination unit, connect with described the second processing unit, for determining the minimum corresponding candidate of distance value
Transmission device is the target transmission device.
Further, the control module includes: the second image identification unit, third processing unit and the first navigation movement
Unit;
Second image identification unit, for identifying target in the target image by visual detection algorithm detection
1 target semanteme points of transmission device;The target semanteme point is to immobilize on the target transmission device, and distinguish
The high point of the property known;
The third processing unit is connect with second image identification unit, for according to the target transmission device
Dimension information, the first spatial position of the target transmission device is calculated;
The first navigation mobile unit, connect, for according to first spatial position with the third processing unit
Navigation is moved to the target transmission device position.
Further, the control module includes: that laser detection unit, fourth processing unit and the second navigation are mobile single
Member;
The laser detection unit obtains each for emitting at detection laser to the stabilizer blade of the target transmission device
The laser coordinate that stabilizer blade is fastened in laser coordinate;
The fourth processing unit is connect with the laser detection unit, for being calculated according to the laser coordinate
The second space position of the target transmission device;
The second navigation mobile unit, connect, for according to the second space position with the fourth processing unit
Navigation is moved to the target transmission device position.
The dispatch control method and system of a kind of transportation robot provided through the invention, can whole process do not need to dispatch
Side participates in that transportation dispatching can be completed, and reduces the retardance of transport task caused by participating in because of dispatching party, and then promote docking
Efficiency and cargo transportation efficiency improve application popularization rate.
Specific embodiment
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, Detailed description of the invention will be compareed below
A specific embodiment of the invention.It should be evident that drawings in the following description are only some embodiments of the invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing, and obtain other embodiments.
To make simplified form, part related to the present invention is only schematically shown in each figure, they are not represented
Its practical structures as product.In addition, there is identical structure or function in some figures so that simplified form is easy to understand
Component only symbolically depicts one of those, or has only marked one of those.Herein, "one" is not only indicated
" only this ", can also indicate the situation of " more than one ".
One embodiment of the present of invention, as depicted in figs. 1 and 2, a kind of dispatch control method of transportation robot 1, comprising:
S100 detects itself working condition and obtains bulk state information;
S200 shooting obtains target image, and target image includes transmission device 2, and is set at transmission device 2, and be used for
Show the indicator light 21 of 2 working condition of transmission device;
S300 analyzes target image to obtain the work state information of transmission device 2;
Specifically, the working condition that transportation robot 1 inquires itself obtains bulk state information, the sheet of transportation robot 1
Body status information includes that transferring robot is in idle and with the to be installed of waiting transmitting device unloading cargo to transportation robot 1
Load state, transferring robot are in idle and have the state to be unloaded for waiting transportation robot 1 to unload cargo to transmitting device,
And transportation robot 1 is in and docks the cargo loading execution state for carrying out cargo loading, or transport with transmitting device
Robot 1 is in docks the cargo unloading execution state for carrying out cargo unloading with transmitting device.In addition, bulk state information
It can also include that obtained available money is analyzed according to own system resource (system resource includes power resources and cpu resource etc.)
Source information, i.e. remaining capacity information and remaining CPU information etc..
Image capture module 11 is provided on transportation robot 1, image capture module 11 is arranged in 1 fuselage of transportation robot
Any one fixed position of front side, so that image capture module 11 can shoot shooting on front side of 1 fuselage of transportation robot
Image in range.Image capture module 11 includes camera, camera, depth camera etc..Transportation robot 1 controls image and adopts
Collecting the shooting of module 11 to obtain includes transmission device 2, and is set at transmission device 2, and for showing 2 working condition of transmission device
Indicator light 21 target image.Formal shooting needs to carry out camera lens adjustment before obtaining target image, to avoid empty burnt phenomenon shadow
Ring the quality of acquired target image.Target image can be picture, be also possible to carry out shot segmentation, obtained figure to video
As frame.After the acquisition of transportation robot 1 obtains target image, image preprocessing is carried out to target image, image preprocessing includes ash
Degreeization processing, binary conversion treatment, filtering processing etc., image preprocessing process are that this is no longer going to repeat them for the prior art.It will
Target image after carrying out image preprocessing is analyzed to obtain the work state information of each transmission device 2.Transmitting device
Work state information include transmission device 2 be in idle and have wait transportation robot 1 unload cargo to transmitting device to
Reception state, transmission device 2 are in idle and have the shape to be sent for waiting transmitting device to unload cargo to transportation robot 1
State and transmitting device are in docks the cargo loading execution state for carrying out cargo loading, Huo Zhechuan with transportation robot 1
Defeated device is in docks the cargo unloading execution state for carrying out cargo unloading with transportation robot 1.
S400 obtains the target transmission device 2 docked with itself according to bulk state information and work state information, analysis
And conveying task type;
S500 obtains the spatial position of target transmission device 2, and navigates and be moved to 2 position of target transmission device;
S600 is docked according to conveying task type with target transmission device 2, and the loading or unloading of cargo is completed.
Specifically, the work state information that transportation robot 1 is obtained according to the bulk state information and analysis of acquisition, into
The docking target i.e. target of conveying task type and each transportation robot 1 that row analysis obtains each transportation robot 1 passes
Send device 2.Then the positioning of transportation robot 1 obtains the location information of itself, and obtains the space bit of target transmission device 2
It sets, path planning is independently carried out according to location information and spatial position and generates mobile route, and then according to the mobile route of generation
It is moved to 2 position of target transmission device, after transportation robot 1 reaches 2 position of target transmission device, with target
Transmission device 2 carries out information exchange, i.e. transportation robot 1 loads and unloads ready trigger signals to target transmission device 2 one,
So that transportation robot 1 is docked between each other with target transmission device 2, complete between target transmission device 2
Cargo loads docking operation or cargo unloads docking operation.
In the present embodiment, need to judge that transportation robot 1 is to receive cargo from transmission device 2, or cargo is shipped to
Transmission device 2, it is therefore desirable to each transportation robot 1 obtains the bulk state information of itself, due to transmission device 2 either
Delivery of goods state is also possible to cargo reception state, and hospital, logistic storage, supermarket or library etc. application scenarios
In might have multiple transmission devices 2, and 2 placement location of multiple transmission devices may be concentrated more, it is therefore desirable to discriminatory analysis
The work state information of each transmission device 2 is obtained, then according to the bulk state information of transportation robot 1 and transmission device 2
Work state information, the target transmission device 2 docked by transportation robot 1 itself matched and searched with itself, and then formulate life
At corresponding cargo transport task, the sky that positioning obtains target transmission device 2 is carried out to the target transmission device 2 docked with itself
Between position, and then navigate be moved to 2 position of target transmission device carry out cargo charge and discharge operations (including cargo load docking
Operation or cargo unload docking operation).The present invention is by transportation robot 1 and 2 organically of transmission device, by linking independently
Determine docking object, and independent navigation is moved to the charge and discharge operations that cargo is completed in destination, is complementary to one another, forms overall efficiency
It is promoted, reaches wider application, effectively solve the problems, such as docking between transportation robot 1 and transmission device 2, and due to complete
Journey, which does not need dispatching party participation, can complete transportation dispatching, reduce the retardance of transport task caused by participating in because of dispatching party,
And then docking efficiency and cargo transportation efficiency are promoted, improve application popularization rate.
Based on previous embodiment, the work state information that S200 is analyzed to obtain transmission device 2 to target image is specific
Comprising steps of
S210 carries out LED status knowledge to target image by visual detection algorithm, using default neural network model
Not;
S220 obtains the work state information of transmission device 2 according to LED status recognition result, analysis;
Wherein, LED status includes instruction lamp on/off state and indicator light shape and color state;Indicator light shape and color state packet
Include indicator light colors state and indicator light shape state.
Specifically, being pre-processed after transportation robot 1 collects target image in real time to target image, with pretreatment
The part in target image including indicator light 21 carries out LED status identification afterwards, to obtain LED status identification knot
Fruit.Preferably, external environment is very big to the Color influences of indicator light 21, possibly even will instruction when carrying out image preprocessing
The color of lamp 21 is as noise filtering, and therefore, indicator light 21 shows transmission device 2 current except through different color states
Except the working condition at moment, can be combined with 21 shape of indicator light and carry out effective supplement, it is effective exclude illumination etc. it is extraneous because
Element and 21 color of indicator light such as decline at the interference of the factors to 21 color of indicator light, realize under the influence of a variety of outsides and internal factor,
The work state information of transmission device 2 can be accurately identified to obtain, so further Lifting Convey machine people 1 according to subject to more
True reliable work state information more precisely finds out in conjunction with bulk state information and passes with the matched target of itself working condition
Device 2 is sent, the accuracy of cargo docking is completed in the robot autonomous scheduling of lifting motion, since the accuracy of cargo docking improves,
And then the probability of mistake docking charge and discharge operations is reduced, to promote docking efficiency and cargo transportation efficiency indirectly.Generally, indicator light
Color state includes red, green, yellow, blue etc., and indicator light shape state includes round, rectangular, triangle etc., is referred to
Show that lamp on/off state includes the illuminating state of indicator light 21 and the OFF state of indicator light 21.Pass through existing visual detection algorithm
LED status identification is carried out, such as it take faster-rcnn as basic structure that default neural network model, which is, back-end network uses
Mobilenetv2 presets the indicator light 21 in neural network model identification positioning target image by this, then carries out indicator light
State recognition.Also for example default neural network model is with R-CNN (or SPP-NET, Fast R-CNN, YOLO, SSD) for base
Plinth structure, back-end network use mobilenetv2 (or mobilenetv1), preset neural network model identification by this and determine
Indicator light 21 in the target image of position, then carries out LED status identification.
In the present embodiment, by visual detection algorithm, determining for indicator light 21 can be carried out using default neural network model
Position identification and the identification of LED status obtain LED status recognition result, and then have by default neural network model
Effect, quick, accurately and efficiently identification obtains LED status recognition result, and detection recognition effect is preferable, is convenient for subsequent transportation machine
Device people 1 analyzes to obtain the work state information of transmission device 2 according to LED status recognition result, so that transportation robot
1 according to bulk state information and work state information, voluntarily matched and searched target transmission device 2, so formulate generate it is corresponding
Cargo transport task precisely, is reliably completed loading operations or unloading operation.Transportation robot 1 is transmitted towards target
2 position of device is mobile, carries out between each other after transportation robot 1 reaches 2 position of target transmission device pair
It connects, and completes loading operations or cargo unloading operation, do not need manually to participate in, reduce costs, it is defeated to improve cargo
Send efficiency.
One embodiment of the present of invention, a kind of dispatch control method of transportation robot 1, comprising:
S100 detects itself working condition and obtains bulk state information;
S200 shooting obtains target image, and target image includes transmission device 2, and is set at transmission device 2, and be used for
Show the indicator light 21 of 2 working condition of transmission device;
S300 analyzes target image to obtain the work state information of transmission device 2;
S301 and remaining transportation robot 1, which carry out the shared working condition for respectively analyzing obtained transmission device 2 of transmission, to be believed
Breath;
S400 obtains the target transmission device 2 docked with itself according to bulk state information and work state information, analysis
And conveying task type;
S500 obtains the spatial position of target transmission device 2, and navigates and be moved to 2 position of target transmission device;
S600 is docked according to conveying task type with target transmission device 2, and the loading or unloading of cargo is completed.
Specifically, this is no longer going to repeat them for part same as the previously described embodiments.The present embodiment is relative to above-mentioned implementation
For example, each machine can carry out information exchange with remaining robot other than itself per capita, and then realize and remaining robot
The work state information for carrying out shared respectively analysis gained transmission device 2, is realized in default scene areas, transportation robot 1 with
Transportation robot 1 shares the work state information respectively obtained between each other, once reach the work shape between transportation robot 1
State information sharing enables to each transportation robot 1 to reduce shooting and obtains all transmission devices 2 and its indicator light around itself
21 target image quantity reduces the working condition for repeating to obtain the same transmission device 2 to analysis in identification synchronization and believes
The probability of breath, and then non-productive work amount is reduced, system resource waste caused by reducing because of non-productive work.Further, since transporter
Work state information between device people 1 is shared, expands the real-time work state information of all transmission devices 2 in wide application scenarios
Tracking, reduce each transportation robot 1 for the blind spot rate of work state information, and then improve each transport in application scenarios
Success rate is docked between robot 1 and transmission device 2, reduces the vacant of transportation robot 1 and transmission device 2 in total system
Rate, and then docking efficiency and cargo transportation efficiency are promoted indirectly.
Based on previous embodiment, S400 obtains docking with itself according to bulk state information and work state information, analysis
Target transmission device 2 and conveying task type specifically include step:
S410 judges that itself is in loading condition or state to be unloaded according to bulk state information;
S420 is in when loading condition when itself, determines that work state information is the corresponding transmission device of state to be sent
2 be target transmission device 2, and determines that conveying task type is that cargo receives type;
S430 determines that work state information is the corresponding transmission device of state to be received when itself is in state to be unloaded
2 be target transmission device 2, and determines that conveying task type is that cargo launches type.
Specifically, transportation robot 1 is analyzed and determined according to the bulk state information of itself, judgement from it is current when
It carves and whether is in loading condition, if transportation robot 1 is at current time to loading condition, stop judging, otherwise
Whether judgement is in state to be unloaded in current time certainly.Can certainly be judgement from current time whether be in
Unloaded state stops judging if transportation robot 1 is in state to be unloaded at current time, and otherwise judgement is from working as
Whether the preceding moment is in loading condition.
Judge that itself is in when loading condition once current transportation robot 1, then according to the work of all transmission devices 2
Status information is judged, judges whether each transmission device 2 is in state to be sent at current time, if current transmission dress
It sets 2 then to stop judging when being in state to be sent at current time, and determines that being in state to be sent at current time corresponds to
Current transmission device 2 be target transmission device 2, and determine transportation robot 1 conveying task type be cargo receive type.
Continue to judge conversely, switching to next transmission device 2, until current transportation robot 1 determines that a transmission device 2 is mesh
Until marking transmission device 2.
When current transportation robot 1 judges that itself is in state to be unloaded, then according to the work of all transmission devices 2
Status information is judged, judges whether each transmission device 2 is in state to be received at current time, if current transmission dress
It sets 2 then to stop judging when being in state to be received at current time, and determines that being in state to be received at current time corresponds to
Current transmission device 2 be target transmission device 2, and determine transportation robot 1 conveying task type be cargo launch type.
Continue to judge conversely, switching to next transmission device 2, until current transportation robot 1 determines that a transmission device 2 is mesh
Until marking transmission device 2.
Illustratively, as shown in Fig. 2, the orientation of several indicator lights 21 includes transversely arranged or longitudinal arrangement,
According to the orientation of indicator light 21, different colours indicator light 21 (shape indicator light 21) will be divided into corresponding in target image
Bright light region, with facilitate to indicator light 21 carry out color identification (or shape recognition).Assuming that two indicator lights 21 are laterally set
It sets, and respectively red indicating light 21 (either round indicator light 21) and green indicator light 21 (or rectangular indicator light 21), and
And setting red indicating light 21 (either round indicator light 21) is bright and when green indicator light 21 (or rectangular indicator light 21) goes out, pass
Device 2 is sent to be in state to be sent, otherwise transmission device 2 is in state to be received.The shooting of transportation robot 1 obtains target figure
Picture, and 21 color of indicator light identification (or shape recognition) and instruction lamp on/off state are carried out to the target image that shooting obtains
Identification.If red indicating light 21 (or round indicator light 21) is bright, it will identify that indicator light 21 is that red colored lamp is (or round
Lamp), so that transportation robot 1 determines that the transmission device 2 of azarin colored lights (or circular lamp) is the target docked with itself
Transmission device 2, and determine that the conveying task type of transportation robot 1 is that cargo receives type.If green indicator light 21 (or side
Shape indicator light 21) it is bright, then it will identify that indicator light 21 is green light (or square front lamp), so that transportation robot 1 determines brilliant green
The transmission device 2 of colored lights (or square front lamp) is the target transmission device 2 docked with itself, and determines transportation robot 1
Conveying task type be cargo launch type.
In the present embodiment, transportation robot 1 can be according to the working condition of itself bulk state information and transmission device 2
Information, the formulation for independently carrying out conveying task generates, to complete to interact docking between transportation robot 1 and transmission device 2
It completes cargo and dispatches transportation work, whole process, which does not need dispatching party participation, can complete transportation dispatching, reduce because dispatching party participates in
The retardance of caused transport task, and then docking efficiency and cargo transportation efficiency are promoted, improve application popularization rate.
Based on previous embodiment, further comprise the steps of:
S401 is when according to bulk state information and work state information, analysis obtains at least two working conditions with itself
When the candidate transmission device 2 to match, the distance between itself and each candidate transmission device 2 value are calculated;
The size of more all distance values of S402 determines that the minimum corresponding candidate transmission device 2 of distance value is target transmission
Device 2.
Specifically, transportation robot 1 may be according to bulk state if including multiple transmission devices 2 at application scenarios
The work state information of information and transmission device 2, analysis obtain at least two and wait for what landing state matched with itself cargo
Candidate transmission device 2, i.e., when transportation robot 1 is in when loading condition, it is understood that there may be at least two are in state to be sent
Transmission device 2, when transportation robot 1 is in state to be unloaded, it is understood that there may be at least two be in state to be received biographies
Send device 2.At this point, transportation robot 1 can be communicated with all with itself working condition matches candidate transmission device 2
Interaction detects signal strength when itself being communicated with each candidate transmission device 2, to calculate to obtain according to signal strength
The distance between itself and each candidate transmission device 2 value.It is, of course, also possible to emit detection signal (such as laser or infrared
Line etc.) to each candidate transmission device 2, after receiving reflected detection signal, according to transmitting detection signal
Launch time and receiving reflects the receiving time of detection signal, and the transfer rate of detection signal calculate
To the distance between transportation robot 1 itself and each candidate transmission device 2 value.In short, no matter which kind of mode, calculates transporter
The method of the distance between device people 1 and each candidate transmission device 2 value is within that scope of the present invention.Transportation robot 1
After the distance between itself and each candidate transmission device 2 value is calculated, all distance values are subjected to size comparison, by distance
Candidate transmission device 2 corresponding to value minimum is used as target transmission device 2.
In the present embodiment, transportation robot 1 is avoided the occurrence of when judging has at least two candidate transmission devices 2, by
In the delay machine phenomenon that can not select and itself carry out docking the appearance of target transmission device 2 for completing cargo handling operation.Pass through
Aforesaid way can independently select the minimum candidate transmission dress of distance value when judging has at least two candidate transmission devices 2
It sets 2 and is used as target transmission device 2, guarantee that each transportation robot 1 is in entire cargo transport process with minimum range progress cargo
Transport, to promote cargo transportation efficiency.
Based on previous embodiment, S500 obtains the spatial position of target transmission device 2, and navigates and be moved to target transmission dress
It sets 2 positions and specifically includes step:
S510 identifies at least four target languages of target transmission device 2 in target image by visual detection algorithm detection
Justice point 22;1 target semanteme points 22 are 2 largest contours vertex of target transmission device, and 1 target semanteme points 22
It is non-coplanar;
The first spatial position of target transmission device 2 is calculated according to the dimension information of target transmission device 2 in S520;
S530 is moved to 2 position of target transmission device according to the navigation of the first spatial position.
Specifically, as shown in Fig. 2, being transmitted and being filled according to determined target after the determining target transmission device 2 of transportation robot 1
It sets 2 corresponding target images and carries out image recognition, since under free scene, the shooting of transportation robot 1 is acquired
Target image in the angle of target transmission device 2 may be different, therefore identify that target transmission fills by visual detection algorithm
After setting 2, the 1 target semanteme points 22 for identifying target transmission device 2 are returned.Semanteme point is that can retouch on transmission device 2
It states, immobilizes and point that identification is high, semanteme point can specifically describe out the position that transmission device 2 is in application scenarios,
Then target semanteme point is then the semantic point on target transmission device 2.Due to the invariance of transmission device 2, then semantic point relative to
The spatial position of application scenarios is then determining, and because semantic point identification is high, then relative to points other on transmission device 2
For, it is easy to return when subsequent image identifies to identify.Semanteme point is more, and semantic point it is more accurate (i.e. semantic point closer to
The largest contours vertex of transmission device 2 is closer) when, the first space where target transmission device 2 is calculated in transportation robot 1
Position position is more accurate.
Transportation robot 1 carries out transportation machines after shooting gets target image, through existing visual detection algorithm
People 1 carries out 1 target semanteme points of 2 fixation and recognition of target transmission device and target transmission device 2 to target image
22 carry out fixation and recognition, and the face feature point that the recurrence of target semanteme point 22 is similar to face returns identification.Such as second is default
It is basic structure that neural network model, which is with faster-rcnn, and back-end network uses mobilenetv2, second default by this
Target transmission device 2 in neural network model identification positioning target image, then advancees to few four target semanteme points 22
Fixation and recognition.Also the such as second default neural network model is with R-CNN (or SPP-NET, Fast R-CNN, YOLO, SSD)
For basic structure, back-end network uses mobilenetv2 (or mobilenetv1), passes through the second default neural network mould
Target transmission device 2 in type identification positioning target image, then advancees to the fixation and recognition of few four target semanteme points 22.By
In target transmission device 2 scale size it is known that using camera calibration as a result, can directly be calculated according to EPnP algorithm
Space coordinate of the 1 target semanteme points 22 relative to image capture module 11, i.e., target transmission device 2 is relative to image
The space coordinate of acquisition module 11.Since the fixation position of transportation robot 1 is arranged in image capture module 11, can then count
Space coordinate of the transportation robot 1 relative to image capture module 11 is calculated, since target transmission device 2 is imaged on Image Acquisition
The pixel coordinate of module 11 it is known that according to the transformational relation between world coordinate system and camera coordinates system carries out that mesh is calculated
Space coordinate of the transmission device 2 on world coordinate system is marked, and the origin due to establishing world coordinate system is it is known that can then obtain
First spatial position of the target transmission device 2 in application scenarios.
Through this embodiment, the default neural network model of building carries out cascade recurrence by foundation structure and back-end network
Target transmission device 2 in recognition target image, then advancees to the fixation and recognition of few four target semanteme points 22, it is this step by step
It returns and knows otherwise, first carry out coarse recurrence identification positioning target object target transmission device 2, then carry out fine return and know
Not Ding Wei target semanteme point 22, constitute one by slightly returning device to the cascade of essence, the phenomenon that then effectively preventing over-fitting,
In identification locating speed and identification locating effect, there is very big promotion.
Above-mentioned building training obtains the training process of the first default neural network model and the second default neural network model
For the prior art, illustratively, predefined frame, which sets the goal, transmission device 2 and defines four target semanteme points 22, obtains mark in advance
Target transmission device 2 is determined and has defined the training sample image of four target semanteme points 22, has been carried out by training sample image
Then training the second default neural network model of building is identified according to the second default neural network model that training obtains.Similarly,
The first default neural network model is obtained according to aforesaid way building training, is no longer described in detail herein.
Based on previous embodiment, S500 obtains the spatial position of target transmission device 2, and navigates and be moved to target transmission dress
It sets 2 positions and specifically includes step:
At S540 transmitting detection laser to the stabilizer blade 23 of target transmission device 2, each stabilizer blade 23 is obtained in laser coordinate system
On laser coordinate;
The second space position of target transmission device 2 is calculated according to laser coordinate by S550;
S560 is moved to 2 position of target transmission device according to the navigation of second space position.
Specifically, sending detection laser to target as shown in Figure 2 by the laser transmitting-receiving equipment being set on transportation robot 1
At the stabilizer blade 23 of transmission device 2.As shown in figure 3, defining Ow-XwYwZw is world coordinate system, Oc-XcYcZc is laser coordinate
System, using Laser emission direction as x-axis, using laser scanning direction as y-axis, x-axis and y-axis constitute scanning surface, perpendicular to scanning surface
For z-axis.Assuming that P point is any one central point of the stabilizer blade 23 of target transmission device 2, then transportation robot 1 can be calculated
Space coordinate of 23 central point of stabilizer blade on world coordinate system.Relationship between laser coordinate system and world coordinate system are as follows:
Wherein, transition matrix R, including laser rotary matrix and laser translation matrix laser.[m, n, 1] is point P
The homogeneous coordinates of laser coordinate of the corresponding laser point p ' in laser coordinate system, [Xw, Yw, Zw, 1] are point P in world coordinates
The homogeneous coordinates of world coordinates in system.Laser rotary matrix and laser translation matrix laser can by multiple groups laser coordinate and
World coordinates carries out calculating acquisition, and for the prior art, this is no longer going to repeat them for this.Building world coordinate system is intended merely to more preferably
Description laser transmitting-receiving equipment, target transmission device 2 spatial position, since laser transmitting-receiving equipment is fixed at transportation machines
On people 1, and transportation robot 1 in the spatial position of application scenarios it is known that due to 2 stabilizer blade 23 of target transmission device central point
231 in the laser coordinate that laser coordinate is fastened it is known that according to the transformational relation between world coordinate system and laser coordinate system, carry out
Space coordinate of the central point 231 of 2 stabilizer blade 23 of target transmission device on world coordinate system is calculated, and due to establishing the world
The origin of coordinate system is it is known that can then obtain second sky of the central point 231 of 2 stabilizer blade 23 of target transmission device in application scenarios
Between position.
Through this embodiment, vertical precise positioning can be carried out in any place of any environment by carrying out positioning using laser,
Adaptable, accuracy rate is high.
One embodiment of the present of invention, as shown in figure 4, a kind of Dispatching Control System of transportation robot 1, comprising: several
A transportation robot 1 and transmission device 2;Transmission device 2 is equipped with the indicator light 21 for showing itself working condition;Each transport
Robot 1 includes: image capture module 11, processing module 13, detection module 12, analysis module 14, control module 15 and executes
Module 16;
Detection module 12 obtains bulk state information for detecting itself working condition;
Image capture module 11 obtains target image for shooting, and target image includes transmission device 2, and is set to and passes
Send the indicator light 21 at device 2;
Processing module 13 is connect with image capture module 11, obtains transmission device 2 for being analyzed target image
Work state information;
Analysis module 14 is connect with processing module 13 and detection module 12 respectively, for according to bulk state information and work
Make status information, analysis obtains the target transmission device 2 docked with itself and conveying task type;
Control module 15 is connect with analysis module 14, for obtaining the spatial position of target transmission device 2, and shifting of navigating
It moves to 2 position of target transmission device;
Execution module 16 is connect with analysis module 14, after being moved to 2 position of target transmission device, according to
Conveying task type carries out docking the loading or unloading for completing cargo with target transmission device 2.
Specifically, the present embodiment is the corresponding Installation practice of above method embodiment, specific effect is referring to the above method
Embodiment, this is no longer going to repeat them.
Based on previous embodiment, processing module 13 includes: the first image identification unit and first processing units;
First image identification unit, for passing through visual detection algorithm, using default neural network model to target image
Carry out LED status identification;
First processing units are connect with the first image identification unit, for analyzing according to LED status recognition result
To the work state information of transmission device 2;
Wherein, LED status includes instruction lamp on/off state and indicator light shape and color state;Indicator light shape and color state packet
Include indicator light colors state and indicator light shape state.
Specifically, the present embodiment is the corresponding Installation practice of above method embodiment, specific effect is referring to the above method
Embodiment, this is no longer going to repeat them.
Based on previous embodiment, each transportation robot 1 further include: wireless communication module;
Wireless communication module is connect with the wireless communication module of remaining transportation robot 1, shared respective for carrying out transmission
Analyze the work state information of obtained transmission device 2.
Based on previous embodiment, analysis module 14 includes: judging unit and the first determination unit;
Judging unit, for judging that itself is in loading condition or state to be unloaded according to bulk state information;
First determination unit, connect with judging unit, for determining that working condition is believed when itself being in when loading condition
Breath is that the corresponding transmission device 2 of state to be sent is target transmission device 2, and determines that conveying task type is that cargo receives class
Type;When itself is in state to be unloaded, determine work state information be the corresponding transmission device 2 of state to be received be target pass
Device 2 is sent, and determines that conveying task type is that cargo launches type.
Specifically, the present embodiment is the corresponding Installation practice of above method embodiment, specific effect is referring to the above method
Embodiment, this is no longer going to repeat them.
Based on previous embodiment, analysis module 14 further include: the second processing unit and the second determination unit;
The second processing unit, for when according to bulk state information and work state information, analysis obtain at least two with
When the candidate transmission device 2 that the working condition of itself matches, the distance between itself and each candidate transmission device 2 are calculated
Value, the size of more all distance values;
Second determination unit, connect with the second processing unit, for determining the minimum corresponding candidate transmission device 2 of distance value
For target transmission device 2.
Specifically, the present embodiment is the corresponding Installation practice of above method embodiment, specific effect is referring to the above method
Embodiment, this is no longer going to repeat them.
Based on previous embodiment, control module 15 includes: the second image identification unit, third processing unit and the first navigation
Mobile unit;
Second image identification unit, for identifying target transmission device in target image by visual detection algorithm detection
21 target semanteme points 22;The target semanteme point 22 is to immobilize on the target transmission device 2, and recognize
The high point of property;
Third processing unit is connect with the second image identification unit, for the dimension information according to target transmission device 2,
The first spatial position of target transmission device 2 is calculated;
First navigation mobile unit, connect with third processing unit, for being moved to mesh according to the navigation of the first spatial position
Mark 2 position of transmission device.
Specifically, the present embodiment is the corresponding Installation practice of above method embodiment, specific effect is referring to the above method
Embodiment, this is no longer going to repeat them.
Based on previous embodiment, control module 15 includes: laser detection unit, fourth processing unit and the second navigation movement
Unit;
Laser detection unit obtains each stabilizer blade 23 for emitting at detection laser to the stabilizer blade 23 of target transmission device 2
In the laser coordinate that laser coordinate is fastened;
Fourth processing unit is connect with laser detection unit, for target transmission device 2 to be calculated according to laser coordinate
Second space position;
Second navigation mobile unit, connect with fourth processing unit, for being moved to mesh according to the navigation of second space position
Mark 2 position of transmission device.
Specifically, the present embodiment is the corresponding Installation practice of above method embodiment, specific effect is referring to the above method
Embodiment, this is no longer going to repeat them.
It should be noted that above-described embodiment can be freely combined as needed.The above is only of the invention preferred
Embodiment, it is noted that for those skilled in the art, in the premise for not departing from the principle of the invention
Under, several improvements and modifications can also be made, these modifications and embellishments should also be considered as the scope of protection of the present invention.