CN116522245A - Pavement type determining method, device, equipment and storage medium - Google Patents
Pavement type determining method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN116522245A CN116522245A CN202310500288.8A CN202310500288A CN116522245A CN 116522245 A CN116522245 A CN 116522245A CN 202310500288 A CN202310500288 A CN 202310500288A CN 116522245 A CN116522245 A CN 116522245A
- Authority
- CN
- China
- Prior art keywords
- model
- road
- road surface
- pavement
- target
- 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
- 239000011159 matrix material Substances 0.000 claims abstract description 85
- 230000007704 transition Effects 0.000 claims abstract description 42
- 239000013598 vector Substances 0.000 claims abstract description 40
- 238000004590 computer program Methods 0.000 claims description 16
- 238000010801 machine learning Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 5
- 238000013145 classification model Methods 0.000 claims description 4
- 239000002689 soil Substances 0.000 description 18
- 239000010426 asphalt Substances 0.000 description 14
- 230000008569 process Effects 0.000 description 10
- 238000004891 communication Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 6
- 238000012805 post-processing Methods 0.000 description 5
- 239000004576 sand Substances 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 210000001503 joint Anatomy 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a pavement type determining method, a pavement type determining device, pavement type determining equipment and a storage medium. The method comprises the following steps: acquiring a state transition matrix between roads, a target observation matrix corresponding to a road classifier model, an initial state probability vector and a model output sequence; constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector; the method and the device for determining the road surface type based on the hidden Markov model and the Viterbi algorithm decode the model output sequence to obtain the target road surface type, and can improve the accuracy of road surface type determination.
Description
Technical Field
The embodiment of the invention relates to the technical field of vehicles, in particular to a pavement type determining method, a pavement type determining device, pavement type determining equipment and a storage medium.
Background
The type of the road surface can influence the running state of the intelligent vehicle, so that the identification of the type of the road surface can provide basis for the control of the intelligent vehicle so as to improve the comfort, the trafficability and the like of the vehicle. However, the action mechanisms of different types of road surfaces and wheels are complex, so that more machine learning algorithms are applied to road surface recognition, but no matter what method is used for recognizing the road surfaces, the recognition results are difficult to be completely accurate, so that the recognition results of the models can shake along with the change of time when the vehicle runs under the same road surface, and if the recognition results of the models are directly applied to the vehicle, the vehicle control is adversely affected.
Disclosure of Invention
The embodiment of the invention provides a pavement type determining method, a device, equipment and a storage medium, which can improve the accuracy of pavement type determination.
According to an aspect of the present invention, there is provided a road surface type determining method including:
acquiring a state transition matrix between roads, a target observation matrix corresponding to a road classifier model, an initial state probability vector and a model output sequence;
constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector;
and decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain the target pavement type.
According to another aspect of the present invention, there is provided a road surface type determining apparatus including:
the acquisition module is used for acquiring a state transition matrix between the roads, a target observation matrix corresponding to the road classifier model, an initial state probability vector and a model output sequence;
the model construction module is used for constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector;
and the pavement type determining module is used for decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain the target pavement type.
According to another aspect of the present invention, there is provided an electronic apparatus including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the road surface type determination method according to any one of the embodiments of the present invention.
According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to execute the road surface type determining method according to any one of the embodiments of the present invention.
According to the embodiment of the invention, the state transition matrix between the road surfaces, the target observation matrix corresponding to the road surface classifier model, the initial state probability vector and the model output sequence are obtained; constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector; decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain a target road surface type, so that when the judgment of the model is fuzzy and the output result is wrong, the post-processing algorithm can quickly correct the output result, and further the accuracy of road surface type determination is improved.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a road surface type determination method in an embodiment of the present invention;
FIG. 2 is a flow chart of a method of determining a target observation matrix in an embodiment of the invention;
FIG. 3 is a flow chart of a post-processing algorithm in an embodiment of the invention;
FIG. 4 is a flow chart of another method of determining a road surface type in an embodiment of the invention;
fig. 5 is a schematic view of a construction of a road surface type determining apparatus in an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device in an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It will be appreciated that prior to using the technical solutions disclosed in the embodiments of the present disclosure, the user should be informed and authorized of the type, usage range, usage scenario, etc. of the personal information related to the present disclosure in an appropriate manner according to the relevant legal regulations.
Example 1
Fig. 1 is a flowchart of a road surface type determining method according to an embodiment of the present invention, where the method may be performed by a road surface type determining device according to an embodiment of the present invention, and the device may be implemented in software and/or hardware, as shown in fig. 1, and the method specifically includes the following steps:
s110, acquiring a state transition matrix between the roads, a target observation matrix corresponding to the road classifier model, an initial state probability vector and a model output sequence.
Specifically, the mode of acquiring the state transition matrix between the road surfaces may be: the state transition matrix between the road surfaces is empirically defined in advance. The mode of acquiring the state transition matrix between the road surfaces can be as follows: acquiring the road surface distribution state of the area where the vehicle is located and the distribution situation of the vehicle on various types of road surfaces; and determining a state transition matrix between the road surfaces according to the road surface distribution state of the area where the vehicle is located and the distribution situation of the vehicle on various types of road surfaces.
Specifically, the method for obtaining the target observation matrix corresponding to the pavement classifier model may be: and (3) pre-establishing a pavement classifier model, acquiring an initial observation matrix, and updating the initial observation matrix according to the output of the pavement classifier model to obtain a target observation matrix.
Specifically, the method for obtaining the initial state probability vector may be: in the initial state, the vehicle obtains insufficient information, and the road surface type of the vehicle in the initial state cannot be determined, so that the initial state probability vector can be defined by referring to the definition mode of the state transition matrix between the road surfaces. In the embodiment of the present invention, the initial state probability vector may be: pi= [ 0.1.0.1.7.0.1 ].
Specifically, the mode of obtaining the model output sequence may be: and obtaining the output of the historical moment pavement classifier model and the output of the current moment pavement classifier model, and constructing a model output sequence according to the output of the historical moment pavement classifier model and the output of the current moment pavement classifier model.
It should be noted that, since the output of the correction model needs to combine the accuracy of the pavement classifier model and the recognition results of the pavement classifier model at several previous moments, a model output result sequence within a period of time needs to be taken, taking the output result of the pavement classifier model for 5 previous times as an example, and adding the output result of the pavement classifier model at the current moment to obtain a model output sequence with a length of 6:
O＝[o1,o2,o3,o4,o5,o6]；
in the formula, o1-o5 are historical classification results output by the model 5 times before the previous moment, and o6 is a classification result output by the model at the current moment and is also a classification result to be corrected.
S120, constructing a hidden Markov model according to the state transition probabilities among the pavements, the target observation matrix corresponding to the pavement classifier model and the initial state probability vector.
Specifically, the method for constructing the hidden markov model according to the state transition probabilities between the pavements, the target observation matrix corresponding to the pavement classifier model, and the initial state probability vector may be as follows: λ= (a, B, pi), where λ is a hidden markov model, a is a state transition matrix between roads, B is a target observation matrix corresponding to a road classifier model, and pi is an initial state probability vector.
S130, decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain the target pavement type.
Specifically, the method for decoding the model output sequence based on the hidden markov model and the Viterbi algorithm may be as follows: decoding the model output sequence based on a hidden Markov model and a Viterbi algorithm to obtain a target state sequence; and determining the last element of the target state sequence as the target pavement type corresponding to the output result to be corrected.
Optionally, obtaining a target observation matrix corresponding to the pavement classifier model includes:
acquiring the probability of each road surface type output by the initial observation matrix and the road surface classifier model in a preset time period;
and updating the initial observation matrix according to the probability of each road surface type output by the road surface classifier model in a preset time period in sequence to obtain a target observation matrix.
The output of the pavement classifier model includes: the classification result and the probability of each road surface type may be, for example, that the output of the road surface classifier model includes: the probability of the soil road is 0.97, the probability of the soil road identified as sand is 0.01, the probability of the asphalt road identified as asphalt road is 0.01, the probability of the ice road identified as ice road is 0.01, and the classification result is: soil path.
Specifically, the method for acquiring the initial observation matrix may be: the method for presetting the initial observation matrix and obtaining the initial observation matrix can be as follows: acquiring the recognition accuracy of the pavement classifier model; and determining an initial observation matrix according to the recognition accuracy of the pavement classifier model.
The probability of each road surface type output in the preset time period may include: the probability of each road surface type output at the historical moment and the probability of each road surface type output at the current moment.
It should be noted that, in the process of identifying the historical model, the accuracy of the model also affects the judgment of the current state, so that the accuracy of the model should be considered in the process of post-processing.
Since the observation matrix in the initial state is only suitable for the overall accuracy of the model, and after a period of time, the matrix has a certain deviation, the observation matrix should be modified and updated in the identification process. When the overall output accuracy of the model is high, the output of the model can be used for defining the observation accuracy of the model after a period of time, the accuracy of model prediction at the previous moment is used for replacing row vectors in the model observation matrix in a certain mode, and the replacement of values in the observation matrix is completed.
For example, when the model output result is [ 0.5.0.3.0.1.1 ] at a certain moment, the model considers that the road surface is a soil road, and the first vector in B is replaced:
when the model output result at the next moment is [ 0.4.0.3.0.1.2 ], the model still considers the road surface to be an earth road, at this time, the first vector in the B is updated, the updating mode can be replaced or the result output by the model is averaged, the earth road is numbered 1 by taking the replacement as an example, and the earth road is obtained after updating the first vector corresponding to the first row vector:
when the result output by the model at the next moment is [ 0.3.0.5.1.0.1 ], the model considers that the road surface is sandy, and the second row vector in the B needs to be replaced:
in the whole post-processing process, the corresponding row vectors in the observation matrix B are continuously replaced by the selected data of 6 states until the required observation matrix B is obtained after the corresponding row vectors in the observation matrix B are replaced based on the result output by the current moment model.
In a specific example, as shown in fig. 2, an initial observation matrix is first determined, a classification result output by a pavement classifier model within a period of time is obtained, a corresponding row vector in the initial observation model is replaced according to the classification result, and a target observation matrix is obtained after the initial observation matrix is replaced based on the classification result output by the model at the current moment.
Optionally, acquiring the initial observation matrix includes:
acquiring the recognition accuracy of the pavement classifier model;
and determining an initial observation matrix according to the recognition accuracy of the pavement classifier model.
Specifically, the manner of obtaining the recognition accuracy of the pavement classifier model may be: and obtaining the probability that the road surface classifier model identifies different types of road surfaces, and determining the model identification accuracy corresponding to the different types of road surfaces according to the probability that the road surface classifier model identifies the different types of road surfaces.
Specifically, the method for determining the initial observation matrix according to the recognition accuracy of the pavement classifier model may be: the method comprises the steps of obtaining accurate model identification probabilities corresponding to different types of road surfaces, optimizing the model identification accuracy corresponding to the different types of road surfaces, and splicing row vectors corresponding to the model identification accuracy corresponding to the optimized different types of road surfaces to obtain an initial observation matrix.
In a specific example, at the initial moment, since the model has insufficient output, the accuracy of model observation of different types of pavement can be defined by means of the performance of the model in the training process, or when the accuracy of the model is high enough, the model accuracy can be simplified to be 1, for example, the probability of model identifying soil road on soil road training set is 0.97, the probability of identifying sand is 0.01, the probability of identifying asphalt road is 0.01, the probability of identifying ice road is 0.01, the model identification accuracy on soil road can be defined to be [0.97 0.01 0.01 0.01], the model identification accuracy under four types of pavement can be simplified to be [ 10 0 0], and row vectors corresponding to the model identification accuracy under four types of pavement are spliced together to form an initial observation matrix B of the model:
optionally, obtaining the model output sequence includes:
acquiring a historical classification result and a classification result to be corrected which are output by a pavement classifier model;
and constructing a model output sequence according to the historical classification result and the classification result to be corrected output by the pavement classifier model.
The historical classification result output by the pavement classifier model may be a classification result output by the model in a period of time, for example, the historical classification result may be: and 5 times of classification results output by the historical moment model are obtained, wherein the classification result to be corrected is the classification result output by the current moment model.
In a specific example, taking the previous 5 output results as an example, adding the model output result at the current moment to obtain a model output sequence with the length of 6: o= [ O1, O2, O3, O4, O5, O6];
in the formula, o1-o5 are classification results output by the model 5 times before the previous moment, and o6 is a classification result output by the model at the current moment and is also a classification result to be corrected.
Optionally, decoding the model output sequence based on a hidden markov model and a Viterbi algorithm to obtain a target road surface type, including:
decoding the model output sequence based on a hidden Markov model and a Viterbi algorithm to obtain a target state sequence;
and determining the last element of the target state sequence as the target pavement type corresponding to the output result to be corrected.
In a specific example, after the state transition matrix between the road surfaces, the target observation matrix corresponding to the road surface classifier model, the initial state probability vector and the model output sequence are obtained, the Viterbi algorithm may be used to post-process the model output result. The observation sequence O is decoded by using the Viterbi algorithm to obtain the most likely state sequence I in the current state, wherein the last element I (6) of the I is the most likely road surface category of the vehicle at the current moment. If the observation sequence o= [1,1,1,1,1,2] output by the model, i= [1, 1] obtained after the Viterbi algorithm decoding, it means that the road classification result output by the model at the present time is sandy (No. 2), and the road most likely in reality is a dirt road (No. 1).
In a specific example, as shown in fig. 3, a post-processing algorithm for a pavement classifier model is started, a state transition matrix a, an observation matrix B and an initial state probability vector are initialized, output of the model in a period of time is obtained, whether the effective length of the model output is met or not is judged, if yes, the observation matrix is updated, after a target observation matrix is obtained, the model output sequence is decoded based on a Viterbi algorithm, and a target pavement type is obtained.
Optionally, before obtaining the state transition probabilities between the pavements, the target observation matrix corresponding to the pavement classifier model, the initial state probability vector and the model output sequence, the method further includes:
acquiring vehicle power characteristic data corresponding to each road surface type;
generating a target sample set according to vehicle dynamic characteristic data corresponding to each road surface type;
establishing a machine learning model;
and training the machine learning model according to the target sample set to obtain a pavement classification model.
Specifically, the manner of acquiring the vehicle power characteristic data corresponding to each road surface type may be: and controlling the test vehicle to respectively run on four road surfaces of the soil road, the sand road, the asphalt road and the ice road, recording dynamic response data of the vehicle, and determining vehicle dynamic characteristic data corresponding to each road surface type according to the dynamic response data of the vehicle.
Specifically, the machine learning model is trained according to the target sample set, and the road classification model may be obtained by: inputting the vehicle power characteristic data in the target sample set into the machine learning model to obtain a predicted road surface type, training parameters of the machine learning model according to an objective function formed by the predicted road surface type and the road surface type corresponding to the vehicle power characteristic data in the target sample set, and returning to execute the operation of inputting the vehicle power characteristic data in the target sample set into the machine learning model to obtain the predicted road surface type until a road surface classification model is obtained.
In a specific example, a real vehicle data acquisition platform is built, and a test vehicle is allowed to travel on four road surfaces of a soil road, a sand road, an asphalt road and an ice road respectively and record dynamic response data of the vehicle, for example: engine torque, wheel speed, vertical acceleration, etc., defining a soil road as a category 1, a sand road as a category 2, an asphalt road as a category 3, and an ice and snow road as a category 4, and then calculating vehicle power characteristic data such as: rolling resistance, wheel speed fluctuation and the like, as well as time domain features and frequency domain features corresponding to the features, constructing a vehicle running data set by utilizing the features, and training a road classifier model with higher accuracy by utilizing the data set.
Optionally, acquiring a state transition matrix between the road surfaces includes:
acquiring the road surface distribution state of the area where the vehicle is located and the distribution situation of the vehicle on various types of road surfaces;
and determining a state transition matrix between the road surfaces according to the road surface distribution state of the area where the vehicle is located and the distribution situation of the vehicle on various types of road surfaces.
Specifically, the probability of transition between various road surfaces in the running process of the vehicle is determined according to the road surface distribution state of the area where the vehicle is located and the distribution situation of the vehicle on various road surfaces. If it is difficult to obtain the state transition probability of each road surface, the state transition probability may be defined empirically, for example, the state transition probability between the same road surface is large, the state transition probability between different road surfaces is small, and the state transition matrix a between the road surfaces may be defined as follows:
the road surface type of the vehicle is rarely changed when the vehicle runs, for example, the road surface type at any moment is an asphalt road when the vehicle runs on a section of asphalt road, the road surface type at any moment before the vehicle runs on a butt joint of the asphalt road and a soil road is an asphalt road when the vehicle runs on a butt joint of the asphalt road and the soil road, and the road surface type at any moment after the vehicle runs on the butt joint of the asphalt road and the soil road is a soil road, so that the running state of the vehicle only changes at the butt joint of the two road surfaces in the process. In summary, the road surface type on which the vehicle is located at a certain moment is more likely to be the same as the road surface on which the vehicle is located at a previous moment. The state transition probability between the same road surfaces is thus greater when the road surfaces are identified. In addition, when the model identifies the road surface type, there is a problem of accuracy, for example, the model may consider that 90% of the probability is soil road, 10% of the probability is asphalt road, and the output identification result is soil road; while the model may consider that there is a 51% likelihood of a soil path and a 49% likelihood of an asphalt path, the output recognition result is still a soil path. The model output results are the same under the two conditions, but the reliability is different, so the embodiment of the invention provides a road surface type determining method, as shown in fig. 4, according to the training of the real vehicle data set, a road surface classifier model is obtained, the characteristics at the historical moment are input into the road surface classifier model to obtain the historical classification result, and the characteristics at the current moment are input into the road surface classifier model to obtain the classification result to be corrected. And constructing a model output sequence according to the historical classification result and the classification result to be corrected output by the pavement classifier model. Obtaining a state transition matrix between roads, a target observation matrix corresponding to a road surface classifier model and an initial state probability vector, constructing a hidden Markov model according to the state transition probability between roads, the target observation matrix corresponding to the road surface classifier model and the initial state probability vector, and decoding an output sequence of the model based on the hidden Markov model and a Viterbi algorithm to obtain an output result, namely the finally determined target road surface type.
According to the technical scheme, a state transition matrix between roads, a target observation matrix corresponding to a road classifier model, an initial state probability vector and a model output sequence are obtained; constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector; and decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain the target road surface type, so that the accuracy of road surface type determination can be improved.
Example two
Fig. 5 is a schematic structural diagram of a road surface type determining device according to an embodiment of the present invention. The present embodiment may be applied to the case of road surface type determination, and the apparatus may be implemented in software and/or hardware, and the apparatus may be integrated in any device that provides a road surface type determination function, as shown in fig. 2, where the road surface type determination apparatus specifically includes: an acquisition module 210, a model construction module 220, and a road surface type determination module 230.
The acquisition module is used for acquiring a state transition matrix between the roads, a target observation matrix corresponding to the road classifier model, an initial state probability vector and a model output sequence;
the model construction module is used for constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector;
and the pavement type determining module is used for decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain the target pavement type.
The product can execute the method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example III
Fig. 6 shows a schematic diagram of the structure of an electronic device 10 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 6, the electronic device 10 includes at least one processor 11, and a memory, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, etc., communicatively connected to the at least one processor 11, in which the memory stores a computer program executable by the at least one processor, and the processor 11 may perform various appropriate actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 may also be stored. The processor 11, the ROM 12 and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
Various components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, etc.; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 11 performs the respective methods and processes described above, such as the road surface type determination method.
In some embodiments, the road surface type determination method may be implemented as a computer program tangibly embodied on a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the road surface type determination method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the road surface type determination method in any other suitable manner (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for carrying out methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.
The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (10)
1. A pavement type determining method, characterized by comprising:
acquiring a state transition matrix between roads, a target observation matrix corresponding to a road classifier model, an initial state probability vector and a model output sequence;
constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector;
and decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain the target pavement type.
2. The method of claim 1, wherein obtaining a target observation matrix corresponding to a pavement classifier model comprises:
acquiring the probability of each road surface type output by the initial observation matrix and the road surface classifier model in a preset time period;
and updating the initial observation matrix according to the probability of each road surface type output by the road surface classifier model in a preset time period in sequence to obtain a target observation matrix.
3. The method of claim 2, wherein obtaining an initial observation matrix comprises:
acquiring the recognition accuracy of the pavement classifier model;
and determining an initial observation matrix according to the recognition accuracy of the pavement classifier model.
4. The method of claim 1, wherein obtaining a model output sequence comprises:
acquiring a historical classification result and a classification result to be corrected which are output by a pavement classifier model;
and constructing a model output sequence according to the historical classification result and the classification result to be corrected output by the pavement classifier model.
5. The method of claim 4, wherein decoding the model output sequence based on a hidden markov model and a Viterbi algorithm to obtain the target road surface type comprises:
decoding the model output sequence based on a hidden Markov model and a Viterbi algorithm to obtain a target state sequence;
and determining the last element of the target state sequence as the target pavement type corresponding to the output result to be corrected.
6. The method of claim 1, further comprising, prior to obtaining the state transition probabilities between the roads, the target observation matrices for the road classifier models, the initial state probability vectors, and the model output sequences:
acquiring vehicle power characteristic data corresponding to each road surface type;
generating a target sample set according to vehicle dynamic characteristic data corresponding to each road surface type;
establishing a machine learning model;
and training the machine learning model according to the target sample set to obtain a pavement classification model.
7. The method of claim 1, wherein obtaining a state transition matrix between roadways comprises:
acquiring the road surface distribution state of the area where the vehicle is located and the distribution situation of the vehicle on various types of road surfaces;
and determining a state transition matrix between the road surfaces according to the road surface distribution state of the area where the vehicle is located and the distribution situation of the vehicle on various types of road surfaces.
8. A road surface type determining apparatus, characterized by comprising:
the acquisition module is used for acquiring a state transition matrix between the roads, a target observation matrix corresponding to the road classifier model, an initial state probability vector and a model output sequence;
the model construction module is used for constructing a hidden Markov model according to the state transition probabilities among the roads, the target observation matrix corresponding to the road classifier model and the initial state probability vector;
and the pavement type determining module is used for decoding the model output sequence based on the hidden Markov model and the Viterbi algorithm to obtain the target pavement type.
9. An electronic device, the electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the road surface type determination method of any one of claims 1-7.
10. A computer readable storage medium storing computer instructions for causing a processor to perform the road surface type determination method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310500288.8A CN116522245A (en) | 2023-05-05 | 2023-05-05 | Pavement type determining method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310500288.8A CN116522245A (en) | 2023-05-05 | 2023-05-05 | Pavement type determining method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116522245A true CN116522245A (en) | 2023-08-01 |
Family
ID=87402663
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310500288.8A Pending CN116522245A (en) | 2023-05-05 | 2023-05-05 | Pavement type determining method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116522245A (en) |
-
2023
- 2023-05-05 CN CN202310500288.8A patent/CN116522245A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109325534B (en) | Semantic segmentation method based on bidirectional multi-scale pyramid | |
| US20220254146A1 (en) | Method for filtering image feature points and terminal | |
| CN113739811A (en) | Method and device for training key point detection model and generating high-precision map lane line | |
| CN116343159B (en) | Unstructured scene passable region detection method, device and storage medium | |
| CN113920158A (en) | Training and traffic object tracking method and device of tracking model | |
| CN116861262B (en) | Perception model training method and device, electronic equipment and storage medium | |
| CN116522245A (en) | Pavement type determining method, device, equipment and storage medium | |
| CN116311893A (en) | Traffic jam time prediction method and device, electronic equipment and storage medium | |
| CN115687764B (en) | Training method of vehicle track evaluation model, vehicle track evaluation method and device | |
| CN113850297B (en) | Road data monitoring method and device, electronic equipment and storage medium | |
| CN113706705B (en) | Image processing method, device, equipment and storage medium for high-precision map | |
| CN115439692A (en) | Image processing method and device, electronic equipment and medium | |
| CN114841338A (en) | Method for training model parameters, decision determination method and device and electronic equipment | |
| CN113901932A (en) | Engineering machinery image recognition method and system fusing artificial fish and particle swarm algorithm | |
| CN116238482A (en) | Parking path planning method, device, equipment and storage medium | |
| CN117523821B (en) | System and method for predicting vehicle multi-mode driving behavior track based on GAT-CS-LSTM | |
| CN117746136A (en) | Pavement disease detection method, neural network training method, device and equipment | |
| CN115099163A (en) | Charge state determining method and device, electronic equipment and storage medium | |
| CN117031512A (en) | Target object tracking method and device and electronic equipment | |
| CN117725834A (en) | Gradient estimation parameter optimization method and device, electronic equipment and storage medium | |
| CN117445795A (en) | Method, device, equipment and medium for controlling turn signal on of vehicle | |
| CN116494969A (en) | Vehicle driving planning method, device, equipment and storage medium | |
| CN112907949A (en) | Traffic anomaly detection method, model training method and device | |
| CN117422363A (en) | Route planning method, device, equipment and storage medium | |
| CN117935208A (en) | Track prediction model training method, track prediction device and vehicle |
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 |