CN110926453A - Obstacle positioning method and system - Google Patents
Obstacle positioning method and system Download PDFInfo
- Publication number
- CN110926453A CN110926453A CN201911072034.0A CN201911072034A CN110926453A CN 110926453 A CN110926453 A CN 110926453A CN 201911072034 A CN201911072034 A CN 201911072034A CN 110926453 A CN110926453 A CN 110926453A
- Authority
- CN
- China
- Prior art keywords
- longitude
- latitude
- coordinates
- coordinate transformation
- transformation model
- 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
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/005—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
Abstract
The application provides a method and a system for positioning an obstacle, comprising the following steps: acquiring longitude and latitude coordinates of a road sampling point, pixel coordinates corresponding to the longitude and latitude coordinates of the road sampling point and pixel coordinates of a road obstacle; training a longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and pixel coordinates corresponding to the longitude and latitude coordinates to obtain a trained longitude and latitude coordinate transformation model; and inputting the pixel coordinates of the road obstacle into the trained longitude and latitude coordinate transformation model to obtain the longitude and latitude coordinates corresponding to the pixel coordinates of the road obstacle in a geodetic coordinate system, and sending the longitude and latitude coordinates to automatic driving equipment or equipment needing global information of the road obstacle for use, so that the problem that the current obstacle positioning information cannot be shared by other automatic driving equipment is solved.
Description
Technical Field
The application relates to the technical field of automatic driving, in particular to a method and a system for positioning an obstacle.
Background
In the fields of automatic driving and assistant driving, obtaining road condition information is a precondition for realizing automatic driving, and at present, road obstacles are mainly positioned in the following way to obtain road condition information, wherein the road condition information includes but is not limited to position information of vehicles on the road, position information of traffic elements such as pedestrians and the like.
One way is to locate road obstacles based on vision, that is, to combine data collected by a camera in an automatic driving device with a map and a GPS, and to determine the position of a vehicle or an obstacle by comparing the data of the camera with data such as the map or the GPS using a probability statistical method.
The other mode is to position the road obstacle based on the laser radar, namely, a detection signal is transmitted to a target obstacle by means of laser radar equipment, and then the received signal reflected by the target obstacle is compared with the transmitted detection signal to obtain the information of the distance, the height and the like of the target obstacle.
Another way is to locate the road obstacle based on ultrasound, i.e. to send an ultrasound signal to the target obstacle by means of an ultrasound device, and to calculate the distance of the obstacle from the time difference of the reception of the ultrasound signal by the receiver.
The above-described obstacle positioning methods are all target obstacle information obtained with respect to the positioning of the autonomous driving apparatus itself, and the target obstacle information cannot be shared by other autonomous driving apparatuses. Secondly, the laser radar is relatively expensive and is also susceptible to weather. Thirdly, since the distance of the obstacle is obtained by ultrasonic positioning, the shape and height information of the obstacle cannot be obtained, and the problem that the obstacle positioning accuracy is not high easily exists.
Disclosure of Invention
The application provides a method and a system for positioning an obstacle, which are used for solving the problem that the existing obstacle positioning information cannot be shared by other automatic driving equipment.
In order to solve the above problem, the present application discloses a method for positioning an obstacle, comprising:
acquiring longitude and latitude coordinates of a road sampling point, pixel coordinates corresponding to the longitude and latitude coordinates of the road sampling point and pixel coordinates of a road obstacle;
training a longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and pixel coordinates corresponding to the longitude and latitude coordinates to obtain a trained longitude and latitude coordinate transformation model;
inputting the pixel coordinates of the road surface barrier into the trained longitude and latitude coordinate transformation model to obtain corresponding longitude and latitude coordinates of the pixel coordinates of the road surface barrier in a geodetic coordinate system;
and sending the longitude and latitude coordinates to automatic driving equipment or equipment needing global information of the road obstacle for use.
Optionally, the step of obtaining the longitude and latitude coordinates of the road surface sampling point and the pixel coordinates corresponding to the longitude and latitude coordinates of the road surface sampling point includes:
arranging a plurality of pavement sampling points on a road;
collecting longitude and latitude coordinates of the road surface sampling points;
acquiring image information of a road sampling point, and acquiring pixel coordinates corresponding to the longitude and latitude coordinates in a camera coordinate system from the image information;
and storing the longitude and latitude coordinates and the corresponding pixel coordinates of the longitude and latitude coordinates in a camera coordinate system.
Optionally, the step of training the longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates to obtain the trained longitude and latitude coordinate transformation model includes:
inputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates into the longitude and latitude coordinate transformation model;
selecting a form and model parameters of the longitude and latitude coordinate transformation model, and giving initial values to the model parameters;
and repeatedly carrying out iterative estimation on the model parameters by adopting an iterative algorithm until the iterative error is smaller than a set threshold value to obtain converged model parameters, and taking the longitude and latitude coordinate transformation model corresponding to the converged model parameters as a trained longitude and latitude coordinate transformation model.
Optionally, the method further includes: and collecting new sample data, reselecting the longitude and latitude coordinate transformation model, training the selected longitude and latitude coordinate transformation model to obtain a trained longitude and latitude coordinate transformation model, and obtaining the longitude and latitude coordinates corresponding to the pixel coordinates of the road obstacle in the geodetic coordinate system by using the trained longitude and latitude coordinate transformation model.
Optionally, the method further includes:
collecting new sample data;
inputting the new sample data into the trained longitude and latitude coordinate transformation model to obtain a predicted value;
and comparing the predicted value with a sample predicted value of a latitude and longitude coordinate transformation model of the sample, and updating model parameters of the trained latitude and longitude coordinate transformation model if the comparison result is smaller than a prediction threshold value.
In order to solve the above problem, the present application also discloses an obstacle positioning system, including:
the data sampling system is used for acquiring longitude and latitude coordinates of a road surface sampling point, pixel coordinates corresponding to the longitude and latitude coordinates of the road surface sampling point and pixel coordinates of a road surface obstacle;
the parameter estimation system is used for training the longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates to obtain the trained longitude and latitude coordinate transformation model;
the longitude and latitude coordinate transformation system is used for inputting the pixel coordinates of the road surface barrier into the trained longitude and latitude coordinate transformation model to obtain corresponding longitude and latitude coordinates of the pixel coordinates of the road surface barrier in a geodetic coordinate system; and sending the longitude and latitude coordinates to automatic driving equipment or equipment needing global information of the road obstacle.
Optionally, the data sampling system includes:
the system comprises an arrangement module, a data processing module and a data processing module, wherein the arrangement module is used for arranging a plurality of road surface sampling points on a road;
the longitude and latitude coordinate module is used for acquiring longitude and latitude coordinates of the road surface sampling point;
the pixel coordinate module is used for acquiring image information of a road surface sampling point and acquiring corresponding pixel coordinates of the longitude and latitude coordinates in a camera coordinate system from the image information;
and the storage module is used for storing the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates in a camera coordinate system.
Optionally, the parameter estimation system includes:
the input module is used for inputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates into the longitude and latitude coordinate transformation model;
the selection module is used for selecting the form and the model parameters of the longitude and latitude coordinate transformation model and endowing the model parameters with initial values;
and the iteration module is used for repeatedly carrying out iterative computation and estimation on the model parameters by adopting an iterative algorithm until the iterative error is smaller than a set threshold value to obtain converged model parameters, and taking the longitude and latitude coordinate transformation model corresponding to the converged model parameters as the trained longitude and latitude coordinate transformation model.
Optionally, the system further includes:
and the updating module is used for collecting new sample data, reselecting the longitude and latitude coordinate transformation model, training the selected longitude and latitude coordinate transformation model to obtain a trained longitude and latitude coordinate transformation model, and obtaining the corresponding longitude and latitude coordinates of the road obstacle pixel coordinates in the geodetic coordinate system by using the trained longitude and latitude coordinate transformation model.
Optionally, the system further includes:
the sample collection module is used for collecting new sample data;
the prediction module is used for inputting the new sample data into the trained longitude and latitude coordinate transformation model to obtain a predicted value;
and the comparison module is used for comparing the predicted value with the sample predicted value of the latitude and longitude coordinate transformation model of the sample, and updating the model parameters of the trained latitude and longitude coordinate transformation model if the comparison result is smaller than a prediction threshold value.
Compared with the prior art, the method has the following advantages:
the method comprises the steps of setting a road surface sampling point, further obtaining longitude and latitude coordinates of the road surface sampling point and pixel coordinates corresponding to the longitude and latitude coordinates of the road surface sampling point, and training a longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates to obtain a trained longitude and latitude coordinate transformation model; thus, the trained longitude and latitude coordinate transformation model is obtained through learning, the pixel coordinates of the road surface barrier are input, and the longitude and latitude coordinates under the geodetic coordinate system corresponding to the pixel coordinates of the barrier, which can be sent to the automatic driving related equipment, can be obtained by calling the trained longitude and latitude coordinate transformation model.
The longitude and latitude coordinates can be quickly calculated from the coordinates of the obstacle through the trained longitude and latitude coordinate transformation model, the calculation amount is very small, and the longitude and latitude coordinates are coordinates under a global coordinate system and can be directly transmitted to automatic driving equipment or other equipment needing global information of the obstacle for use, for example: high-definition maps.
Of course, it is not necessary for any product to achieve all of the above-described advantages at the same time for practicing the present application.
Drawings
Fig. 1 is a flowchart of an obstacle locating method according to an embodiment of the present application;
FIG. 2 is an example of a data application according to an embodiment of the present application;
fig. 3 is a block diagram of an obstacle locating system according to an embodiment of the present application.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present application more comprehensible, the present application is described in further detail with reference to the accompanying drawings and the detailed description.
Referring to fig. 1, a flowchart of an obstacle positioning method according to an embodiment of the present application is shown, which specifically includes:
step 101: and acquiring longitude and latitude coordinates of the road surface sampling point, pixel coordinates corresponding to the longitude and latitude coordinates of the road surface sampling point and pixel coordinates of the road surface barrier.
As one of the implementation, step 101 includes the following sub-steps:
substep 1011: a plurality of pavement sampling points are arranged on a road.
In practical application, the arrangement of sampling points can influence the precision of estimating the longitude and latitude positions by the longitude and latitude coordinate transformation model, so that the more the number of the sampling points is, the better the sampling points are, meanwhile, the whole sampling area is covered by attention, and a sampling designer needs to determine the number and the positions of the road surface sampling points according to the actual condition of a road.
In the vehicle-road cooperative automatic driving technology, a road test sensing unit arranged on a road continuously monitors a fixed road area, and long-time continuous analysis processing is carried out on road conditions, so that more precise and more accurate information of the road conditions in the road area can be obtained.
For example: the road range monitored by the drive test sensing unit is L W (unit: meter), a sampling point is marked by red paint every 2 meters in the length and width directions, and the radius of the mark is about 4 centimeters.
It should be noted that the above examples are merely examples, and in practical applications, the adjustment may be performed again according to the size of the road range, and the application is not limited in particular.
Substep 1012: and collecting longitude and latitude coordinates of the road surface sampling points.
And aligning the sampling points of the road surface by using the handheld longitude and latitude detection device at the sampling points of the road surface, measuring the longitude and latitude coordinates of the sampling points, finally obtaining the longitude and latitude coordinates of a plurality of sampling points of the road surface, and finally forming a longitude and latitude coordinate list corresponding to the sampling points of the road surface.
In general, the latitude and longitude coordinates are accurate to the 7 th digit after the decimal point.
In practical application, a longitude and latitude coordinate list corresponding to the road surface sampling point is formed, and the method specifically comprises the following steps: selecting one of the road sampling points, collecting longitude and latitude coordinates of the sampling point by using a longitude and latitude detection device, moving to the next sampling point after the sampling is finished, collecting the longitude and latitude coordinates of the road sampling point until all the sampling points are collected, and then forming a longitude and latitude list corresponding to the road sampling point.
In the embodiment, the longitude and latitude coordinates are measured by using the longitude and latitude detection device in a point-by-point mode, only one longitude and latitude measurement device is needed, the cost for purchasing or leasing the longitude and latitude measurement instrument can be reduced, and the sampled data can be more accurate by using the high-precision longitude and latitude measurement device.
Substep 1013: and acquiring image information of a road sampling point, and acquiring pixel coordinates corresponding to the longitude and latitude coordinates in a camera coordinate system from the image information.
After the longitude and latitude coordinates are obtained, acquiring image information of a road sampling point by arranging a road monitoring camera or a terminal with a camera shooting function, and positioning corresponding pixel coordinates of the longitude and latitude coordinates in a camera coordinate system in the image information.
In practical application, the pixel coordinates corresponding to the longitude and latitude coordinates in the camera coordinate system may be obtained according to a preset relationship between the longitude and latitude coordinates and the pixel coordinates, or the pixel coordinates corresponding to the longitude and latitude coordinates in the camera coordinate system may be obtained by using a preset model, or may be obtained by using other methods, which is not limited in this application.
Substep 1014: and storing the longitude and latitude coordinates and the corresponding pixel coordinates of the longitude and latitude coordinates in a camera coordinate system.
For the corresponding relation of the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude, one line can be adopted to store one sampling point data, and the data format is longitude, latitude and pixel coordinates, and can also be stored in other modes.
And storing the longitude and latitude coordinates and the corresponding pixel coordinates of the longitude and latitude coordinates in a camera coordinate system into a storage device or directly sending the longitude and latitude coordinates to a target device for use in other links, thereby providing support for automatic driving path planning and decision making.
The target device may be an automatic driving device, a device that needs global information of a road obstacle, or other devices, and the application is not particularly limited.
It should be noted that, in practical applications, after all operations of sub-step 1012 are completed, sub-step 1013 is executed, or after one latitude and longitude coordinate is collected by sub-step 1012, sub-step 1013 is executed, which is not limited in this application.
Referring to fig. 2, an example of a data sampling application is shown, which specifically includes the following steps:
step 201: a plurality of pavement sampling points disposed on a roadway is determined.
Step 202: a sample point is selected.
Step 203: and acquiring longitude and latitude coordinates of the sampling points.
Step 204: and shooting image information of the sampling points.
Step 205: and calculating the pixel coordinates of the longitude and latitude coordinates in the camera coordinate system in the image information.
Step 206: it is determined whether the sampling is complete, if so, step 207 is performed, and if not, step 208 is performed.
Step 207: and outputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates in a camera coordinate system.
Step 208: move to the next sample point and repeat steps 203-206.
The data sampling provides input data for a subsequent longitude and latitude coordinate transformation model, so that the key of a sampling system is to ensure the correctness and the effectiveness of the sampled data, and meanwhile, the data sampling ensures the corresponding relation between the acquired pixel coordinate and the longitude and latitude coordinate.
As another implementation, step 101 includes the following sub-steps:
the pixel coordinates of the road surface barrier are collected through a camera or a terminal with a camera shooting function, and the pixel coordinates of the road surface barrier are stored in a storage device or sent to a target device.
When the camera shoots, images containing and not containing the obstacle exist, and then an obstacle detection algorithm is operated to obtain the pixel coordinates of the obstacle in a camera coordinate system.
Step 102: and training the longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates to obtain the trained longitude and latitude coordinate transformation model.
And inputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates into a longitude and latitude coordinate transformation model to obtain a longitude and latitude coordinate transformation model, and calling the trained longitude and latitude coordinate transformation model to obtain the longitude and latitude coordinates corresponding to the pixel coordinates of the road surface barrier in a geodetic coordinate system.
In practical application, for image information including urban roads, the pixel point coordinates of each road surface have longitude and latitude coordinates corresponding to the pixel point coordinates, and one-to-one correspondence exists between the pixel point coordinates and the longitude and latitude coordinates, so that a mathematical model of the correspondence can be established according to a group of sampling points, and finally the longitude and latitude of the pixel coordinates can be calculated.
In practical application, the trained longitude and latitude coordinate transformation model can be described by a linear or nonlinear coordinate transformation model.
The linear relationship may use a straight-line function, such as: y ═ ax + b, the non-linear relationship may use a parabolic function, such as: y is ax2+ bx + c, other functions may be employed, and the above functions are used as examples only.
The longitude and latitude coordinate transformation model specifically works as follows: inputting sampling data in a longitude and latitude coordinate transformation model, wherein the sampling data comprises: after sampling data are input and a longitude and latitude coordinate transformation model is selected, estimating a model parameter value by adopting an iterative method according to the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates: firstly initializing parameters, calculating longitude and latitude estimated values of pixel values of each sampling point, then calculating errors between the longitude and latitude sampling values and the estimated values, then continuously repeating the previous processes by using error adjusting parameters until the errors are converged to an acceptable degree, and finally storing the model parameters.
Specifically, step 102 includes the following sub-steps:
substep 1021: and inputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates into the longitude and latitude coordinate transformation model.
Substep 1022: and selecting a form and model parameters of the longitude and latitude coordinate transformation model, and giving initial values to the model parameters.
The form of the longitude and latitude coordinate transformation model can be a polynomial order form, and can also be other forms, and the application is not particularly limited.
Substep 1023: and repeatedly carrying out iterative estimation on the model parameters by adopting an iterative algorithm until the iterative error is smaller than a set threshold value to obtain converged model parameters, and taking the longitude and latitude coordinate transformation model corresponding to the converged model parameters as a trained longitude and latitude coordinate transformation model.
In practical application, a Gaussian-Seidel iteration estimation model parameter can be adopted, the error of a sample is calculated, whether the error is smaller than a set threshold value or not is judged, convergence is indicated if the error is smaller than the set threshold value, a converged model parameter is obtained, a longitude and latitude coordinate transformation model corresponding to the converged model parameter is obtained and serves as a trained longitude and latitude coordinate transformation model, and if the error is larger than the set threshold value, the Gaussian-Seidel iteration estimation model parameter is continued until convergence.
Step 103: and inputting the pixel coordinates of the road surface barrier into the trained longitude and latitude coordinate transformation model to obtain the corresponding longitude and latitude coordinates of the pixel coordinates of the road surface barrier in a geodetic coordinate system.
The pixel coordinates of the road surface barrier are input into the trained longitude and latitude coordinate transformation model, so that the longitude and latitude coordinates corresponding to the pixel coordinates of the road surface barrier in a geodetic coordinate system can be obtained, and the longitude and latitude coordinates are sent to other automatic driving related equipment.
In practical application, the coordinates of the obstacle in the camera coordinate system are converted into longitude and latitude coordinates under a geodetic coordinate system WGS-84, so that the real position information of the obstacle can be known and used by automatic driving related equipment, and support is provided for automatic driving path planning and decision making.
For example: the trained longitude and latitude coordinate transformation model is that y is 2x +1, wherein x is the pixel coordinate of the obstacle, y is the longitude and latitude coordinate, and if the pixel coordinate of the obstacle is 1, the obtained longitude and latitude coordinate is 3.
For example: the trained longitude and latitude coordinate transformation model is that y is x2+2x +1, where x is the pixel coordinate of the obstacle, y is the longitude and latitude coordinate, and if the pixel coordinate of the obstacle is 1, the obtained longitude and latitude coordinate is 4.
Step 104: and sending the longitude and latitude coordinates to automatic driving equipment or equipment needing global information of the road obstacle for use.
Further, the method further comprises: and updating the trained longitude and latitude coordinate transformation model.
The updating of the trained longitude and latitude coordinate transformation model comprises two modes, wherein one mode is as follows: only updating the model parameters of the trained longitude and latitude coordinate transformation model, and the other mode is as follows: and replacing the model and then estimating the mode parameters again.
The updating of the model parameters of the trained longitude and latitude coordinate transformation model can be realized in the following ways:
collecting new sample data; inputting the new sample data into the trained longitude and latitude coordinate transformation model to obtain a predicted value; and comparing the predicted value with a sample predicted value of a parameter estimation model of a sample, and updating the model parameters of the trained longitude and latitude coordinate transformation model if the comparison result is smaller than a prediction threshold value.
Wherein, the model replacement can be realized by the following modes:
and (3) reselecting the longitude and latitude coordinate transformation model according to the collected new sample data, training the selected longitude and latitude coordinate transformation model to obtain a trained longitude and latitude coordinate transformation model, and obtaining the longitude and latitude coordinates corresponding to the pixel coordinates of the road obstacle in the geodetic coordinate system by using the trained longitude and latitude coordinate transformation model.
As one implementation manner, the longitude and latitude coordinate transformation model may be a neural network model, and after new sample data is collected, the neural network model is selected and trained, so as to obtain a trained neural network model, and the neural network model is used to obtain the longitude and latitude coordinates corresponding to the pixel coordinates of the road obstacle in the geodetic coordinate system.
The neural network model may be a three-layer neural network or a four-layer neural network, and the application is not particularly limited.
The new sample data may be the newly acquired longitude and latitude coordinates of the road surface sampling point, the pixel coordinates corresponding to the longitude and latitude coordinates, and the like, and may also be other data, which is not specifically limited in this application.
In this embodiment, by setting a road surface sampling point, a longitude and latitude coordinate of the road surface sampling point and a pixel coordinate corresponding to the longitude and latitude coordinate of the road surface sampling point are obtained, and a longitude and latitude coordinate transformation model is trained according to the longitude and latitude coordinate and the pixel coordinate corresponding to the longitude and latitude coordinate, so as to obtain a trained longitude and latitude coordinate transformation model.
Thus, the trained longitude and latitude coordinate transformation model is obtained through learning, the mapping relation between the longitude and latitude coordinates of the sampling point and the corresponding pixel coordinates is expressed, the pixel coordinates of the road surface obstacle are input, the trained longitude and latitude coordinate transformation model is called to obtain the longitude and latitude coordinates under the geodetic coordinate system corresponding to the obstacle pixel coordinates which can be sent to the automatic driving related equipment, the longitude and latitude coordinates can be rapidly calculated from the obstacle coordinates through the trained longitude and latitude coordinate transformation model, the calculated amount is very small, and the longitude and latitude coordinates are the coordinates under the global coordinate system and can be directly transmitted to the automatic driving equipment or other equipment needing the global information of the obstacle for use, for example: high-definition maps.
It should be noted that the foregoing method embodiments are described as a series of acts or combinations for simplicity in explanation, but it should be understood by those skilled in the art that the present application is not limited by the order of acts described, as some steps may occur in other orders or concurrently depending on the application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are presently preferred and that no particular act is required to implement the invention.
Based on the description of the above method embodiment, the present application also provides a corresponding system embodiment to implement the content described in the above method embodiment.
Referring to fig. 3, which shows a structure diagram of an obstacle positioning system according to an embodiment of the present application, specifically, the structure diagram includes:
the data sampling system 301 is configured to acquire longitude and latitude coordinates of a road surface sampling point, pixel coordinates corresponding to the longitude and latitude coordinates of the road surface sampling point, and pixel coordinates of a road surface obstacle.
And the parameter estimation system 302 is used for training the longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates to obtain the trained longitude and latitude coordinate transformation model.
And the longitude and latitude coordinate transformation system 303 is used for inputting the pixel coordinates of the road obstacle into the trained longitude and latitude coordinate transformation model to obtain longitude and latitude coordinates corresponding to the pixel coordinates of the road obstacle, and sending the longitude and latitude coordinates to automatic driving equipment or equipment needing global information of the road obstacle for use.
As one implementation manner, a sending module may be disposed in the longitude and latitude coordinate transformation system 303, and configured to send the longitude and latitude coordinates to an automatic driving device or a device that needs global information of a road obstacle.
Optionally, the data sampling system includes:
the arrangement module is used for arranging a plurality of road surface sampling points on a road.
And the longitude and latitude coordinate module is used for acquiring longitude and latitude coordinates of the road surface sampling point.
And the pixel coordinate module is used for acquiring the image information of the road surface sampling point and acquiring the pixel coordinate corresponding to the longitude and latitude coordinate from the image information.
And the storage module is used for storing the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates.
Optionally, the parameter estimation system includes:
and the input module is used for inputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates into the longitude and latitude coordinate transformation model.
And the selection module is used for selecting the form and the model parameters of the longitude and latitude coordinate transformation model and endowing the model parameters with initial values.
And the iteration module is used for repeatedly carrying out iterative computation and estimation on the model parameters by adopting an iterative algorithm until the iterative error is smaller than a set threshold value, obtaining converged model parameters and taking the longitude and latitude coordinate transformation model corresponding to the converged model parameters as the trained longitude and latitude coordinate transformation model.
Optionally, the apparatus further comprises:
and the sample collection module is used for collecting new sample data.
And the prediction module is used for inputting the new sample data into the trained longitude and latitude coordinate transformation model to obtain a predicted value.
And the comparison module is used for comparing the predicted value with the sample predicted value of the parameter estimation model of the sample, and updating the trained longitude and latitude coordinate transformation model if the comparison result is smaller than a prediction threshold value.
In this embodiment, by setting a road surface sampling point, a longitude and latitude coordinate of the road surface sampling point and a pixel coordinate corresponding to the longitude and latitude coordinate of the road surface sampling point are obtained, and a longitude and latitude coordinate transformation model is trained according to the longitude and latitude coordinate and the pixel coordinate corresponding to the longitude and latitude coordinate, so as to obtain a trained longitude and latitude coordinate transformation model.
Thus, the trained longitude and latitude coordinate transformation model is obtained through learning, the mapping relation between the longitude and latitude coordinates of the sampling point and the corresponding pixel coordinates is expressed, the pixel coordinates of the road surface obstacle are input, the trained longitude and latitude coordinate transformation model is called to obtain the longitude and latitude coordinates under the geodetic coordinate system corresponding to the obstacle pixel coordinates which can be sent to the automatic driving related equipment, the longitude and latitude coordinates can be rapidly calculated from the obstacle coordinates through the trained longitude and latitude coordinate transformation model, the calculated amount is very small, and the longitude and latitude coordinates are the coordinates under the global coordinate system and can be directly transmitted to the automatic driving equipment or other equipment needing the global information of the obstacle for use, for example: high-definition maps.
For the above apparatus embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the illustrated method embodiment.
The embodiments in the present specification are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
As is readily imaginable to the person skilled in the art: any combination of the above embodiments is possible, and thus any combination between the above embodiments is an embodiment of the present application, but the present disclosure is not necessarily detailed herein for reasons of space.
In this application, "component," "device," "system," and the like are intended to refer to a computer-related entity, either hardware, a combination of hardware and software, or software in execution. In particular, for example, a component can be, but is not limited to being, a process running on a processor, an object, an executable, a thread of execution, a program, and/or a computer. Also, an application or script running on a server, or a server, can be a component. One or more components can reside within a process and/or thread of execution and a component can be localized on one computer and/or distributed between two or more computers and can be run by various computer-readable media. The components may also communicate by way of local and/or remote processes in accordance with a signal having one or more data packets, e.g., signals from data interacting with another component in a local system, distributed system, and/or across a network of the internet with other systems by way of the signal.
Finally, it should also be noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
Further, the word "and/or" above means that the relation "and" or "is included herein, wherein: if the scheme A and the scheme B are in an 'and' relationship, the method indicates that the scheme A and the scheme B can be simultaneously included in a certain embodiment; if the scheme a and the scheme B are in an or relationship, this means that in some embodiment, the scheme a may be included separately, or the scheme B may be included separately.
While the preferred embodiments of the present application have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all alterations and modifications as fall within the scope of the application.
The method and system for positioning the obstacle provided by the present application are introduced in detail, and a specific example is applied in the text to explain the principle and the implementation of the present application, and the description of the above embodiment is only used to help understand the method and the core idea of the present application; meanwhile, for a person skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.
Claims (10)
1. A method of obstacle location, comprising:
acquiring longitude and latitude coordinates of a road sampling point, pixel coordinates corresponding to the longitude and latitude coordinates of the road sampling point and pixel coordinates of a road obstacle;
training a longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and pixel coordinates corresponding to the longitude and latitude coordinates to obtain a trained longitude and latitude coordinate transformation model;
inputting the pixel coordinates of the road surface barrier into the trained longitude and latitude coordinate transformation model to obtain corresponding longitude and latitude coordinates of the pixel coordinates of the road surface barrier in a geodetic coordinate system;
and sending the longitude and latitude coordinates to automatic driving equipment or equipment needing global information of the road obstacle for use.
2. The method of claim 1, wherein the step of obtaining the longitude and latitude coordinates of the road surface sampling point and the pixel coordinates corresponding to the longitude and latitude coordinates of the road surface sampling point comprises:
arranging a plurality of pavement sampling points on a road;
collecting longitude and latitude coordinates of the road surface sampling points;
acquiring image information of a road sampling point, and acquiring pixel coordinates corresponding to the longitude and latitude coordinates in a camera coordinate system from the image information;
and storing the longitude and latitude coordinates and the corresponding pixel coordinates of the longitude and latitude coordinates in a camera coordinate system.
3. The method of claim 1, wherein the step of training the latitude and longitude coordinate transformation model according to the latitude and longitude coordinates and the pixel coordinates corresponding to the latitude and longitude coordinates to obtain the trained latitude and longitude coordinate transformation model comprises:
inputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates into the longitude and latitude coordinate transformation model;
selecting a form and model parameters of the longitude and latitude coordinate transformation model, and giving initial values to the model parameters;
and repeatedly carrying out iterative estimation on the model parameters by adopting an iterative algorithm until the iterative error is smaller than a set threshold value to obtain converged model parameters, and taking the longitude and latitude coordinate transformation model corresponding to the converged model parameters as a trained longitude and latitude coordinate transformation model.
4. The method of claim 1, further comprising: and collecting new sample data, reselecting the longitude and latitude coordinate transformation model, training the selected longitude and latitude coordinate transformation model to obtain a trained longitude and latitude coordinate transformation model, and obtaining the longitude and latitude coordinates corresponding to the pixel coordinates of the road obstacle in the geodetic coordinate system by using the trained longitude and latitude coordinate transformation model.
5. The method according to any one of claims 1-3, further comprising:
collecting new sample data;
inputting the new sample data into the trained longitude and latitude coordinate transformation model to obtain a predicted value;
and comparing the predicted value with a sample predicted value of a latitude and longitude coordinate transformation model of the sample, and updating model parameters of the trained latitude and longitude coordinate transformation model if the comparison result is smaller than a prediction threshold value.
6. An obstacle locating system, comprising:
the data sampling system is used for acquiring longitude and latitude coordinates of a road surface sampling point, pixel coordinates corresponding to the longitude and latitude coordinates of the road surface sampling point and pixel coordinates of a road surface obstacle;
the parameter estimation system is used for training the longitude and latitude coordinate transformation model according to the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates to obtain the trained longitude and latitude coordinate transformation model;
the longitude and latitude coordinate transformation system is used for inputting the pixel coordinates of the road surface barrier into the trained longitude and latitude coordinate transformation model to obtain corresponding longitude and latitude coordinates of the pixel coordinates of the road surface barrier in a geodetic coordinate system; and sending the longitude and latitude coordinates to automatic driving equipment or equipment needing global information of the road obstacle.
7. The obstacle positioning system of claim 6, wherein the data sampling system comprises:
the system comprises an arrangement module, a data processing module and a data processing module, wherein the arrangement module is used for arranging a plurality of road surface sampling points on a road;
the longitude and latitude coordinate module is used for acquiring longitude and latitude coordinates of the road surface sampling point;
the pixel coordinate module is used for acquiring image information of a road surface sampling point and acquiring corresponding pixel coordinates of the longitude and latitude coordinates in a camera coordinate system from the image information;
and the storage module is used for storing the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates in a camera coordinate system.
8. The obstacle positioning system of claim 6, wherein the parameter estimation system comprises:
the input module is used for inputting the longitude and latitude coordinates and the pixel coordinates corresponding to the longitude and latitude coordinates into the longitude and latitude coordinate transformation model;
the selection module is used for selecting the form and the model parameters of the longitude and latitude coordinate transformation model and endowing the model parameters with initial values;
and the iteration module is used for repeatedly carrying out iterative computation and estimation on the model parameters by adopting an iterative algorithm until the iterative error is smaller than a set threshold value to obtain converged model parameters, and taking the longitude and latitude coordinate transformation model corresponding to the converged model parameters as the trained longitude and latitude coordinate transformation model.
9. The obstacle positioning system of claim 6, further comprising:
and the updating module is used for collecting new sample data, reselecting the longitude and latitude coordinate transformation model, training the selected longitude and latitude coordinate transformation model to obtain a trained longitude and latitude coordinate transformation model, and obtaining the corresponding longitude and latitude coordinates of the road obstacle pixel coordinates in the geodetic coordinate system by using the trained longitude and latitude coordinate transformation model.
10. An obstacle positioning system according to any of claims 6-8, wherein the system further comprises:
the sample collection module is used for collecting new sample data;
the prediction module is used for inputting the new sample data into the trained longitude and latitude coordinate transformation model to obtain a predicted value;
and the comparison module is used for comparing the predicted value with the sample predicted value of the latitude and longitude coordinate transformation model of the sample, and updating the model parameters of the trained latitude and longitude coordinate transformation model if the comparison result is smaller than a prediction threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911072034.0A CN110926453A (en) | 2019-11-05 | 2019-11-05 | Obstacle positioning method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911072034.0A CN110926453A (en) | 2019-11-05 | 2019-11-05 | Obstacle positioning method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110926453A true CN110926453A (en) | 2020-03-27 |
Family
ID=69852400
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911072034.0A Pending CN110926453A (en) | 2019-11-05 | 2019-11-05 | Obstacle positioning method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110926453A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113804100A (en) * | 2020-06-11 | 2021-12-17 | 华为技术有限公司 | Method, device, equipment and storage medium for determining space coordinates of target object |
| CN115334247A (en) * | 2022-10-11 | 2022-11-11 | 齐鲁空天信息研究院 | Camera module calibration method, visual positioning method and device and electronic equipment |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101255461B1 (en) * | 2011-09-28 | 2013-04-16 | 한국건설기술연구원 | Position Measuring Method for street facility |
| CN104123695A (en) * | 2014-06-23 | 2014-10-29 | 唐文兴 | Method for realizing coordinate conversion |
| CN105928504A (en) * | 2016-04-14 | 2016-09-07 | 北京小米移动软件有限公司 | Barrier positioning method, device and system |
| CN106530218A (en) * | 2016-10-28 | 2017-03-22 | 浙江宇视科技有限公司 | Coordinate conversion method and apparatus |
| CN107845060A (en) * | 2017-10-31 | 2018-03-27 | 广东中星电子有限公司 | Geographical position and corresponding image position coordinates conversion method and system |
| CN107966155A (en) * | 2017-12-25 | 2018-04-27 | 北京地平线信息技术有限公司 | Object positioning method, object positioning system and electronic equipment |
| KR101941654B1 (en) * | 2018-09-28 | 2019-01-24 | (주)신한항업 | Image editing system for gps coordinate on digital map searching position and link information |
| CN109657638A (en) * | 2018-12-28 | 2019-04-19 | 百度在线网络技术（北京）有限公司 | Barrier localization method, device and terminal |
| CN110148185A (en) * | 2019-05-22 | 2019-08-20 | 北京百度网讯科技有限公司 | Determine method, apparatus, electronic equipment and the storage medium of coordinate system conversion parameter |
| CN110174093A (en) * | 2019-05-05 | 2019-08-27 | 腾讯科技（深圳）有限公司 | Localization method, device, equipment and computer readable storage medium |
| CN110197461A (en) * | 2019-06-06 | 2019-09-03 | 上海木木聚枞机器人科技有限公司 | A kind of coordinate transformation relation determines method, apparatus, equipment and storage medium |
| CN110208739A (en) * | 2019-05-29 | 2019-09-06 | 北京百度网讯科技有限公司 | Assist method, apparatus, equipment and the computer readable storage medium of vehicle location |
-
2019
- 2019-11-05 CN CN201911072034.0A patent/CN110926453A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101255461B1 (en) * | 2011-09-28 | 2013-04-16 | 한국건설기술연구원 | Position Measuring Method for street facility |
| CN104123695A (en) * | 2014-06-23 | 2014-10-29 | 唐文兴 | Method for realizing coordinate conversion |
| CN105928504A (en) * | 2016-04-14 | 2016-09-07 | 北京小米移动软件有限公司 | Barrier positioning method, device and system |
| CN106530218A (en) * | 2016-10-28 | 2017-03-22 | 浙江宇视科技有限公司 | Coordinate conversion method and apparatus |
| CN107845060A (en) * | 2017-10-31 | 2018-03-27 | 广东中星电子有限公司 | Geographical position and corresponding image position coordinates conversion method and system |
| CN107966155A (en) * | 2017-12-25 | 2018-04-27 | 北京地平线信息技术有限公司 | Object positioning method, object positioning system and electronic equipment |
| KR101941654B1 (en) * | 2018-09-28 | 2019-01-24 | (주)신한항업 | Image editing system for gps coordinate on digital map searching position and link information |
| CN109657638A (en) * | 2018-12-28 | 2019-04-19 | 百度在线网络技术（北京）有限公司 | Barrier localization method, device and terminal |
| CN110174093A (en) * | 2019-05-05 | 2019-08-27 | 腾讯科技（深圳）有限公司 | Localization method, device, equipment and computer readable storage medium |
| CN110148185A (en) * | 2019-05-22 | 2019-08-20 | 北京百度网讯科技有限公司 | Determine method, apparatus, electronic equipment and the storage medium of coordinate system conversion parameter |
| CN110208739A (en) * | 2019-05-29 | 2019-09-06 | 北京百度网讯科技有限公司 | Assist method, apparatus, equipment and the computer readable storage medium of vehicle location |
| CN110197461A (en) * | 2019-06-06 | 2019-09-03 | 上海木木聚枞机器人科技有限公司 | A kind of coordinate transformation relation determines method, apparatus, equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| AMAZON: ""Retraining Models on New Data"", 《AMAZON MACHINE LEARNING DEVELOPER GUIDE》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113804100A (en) * | 2020-06-11 | 2021-12-17 | 华为技术有限公司 | Method, device, equipment and storage medium for determining space coordinates of target object |
| CN115334247A (en) * | 2022-10-11 | 2022-11-11 | 齐鲁空天信息研究院 | Camera module calibration method, visual positioning method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10621696B2 (en) | Deep convolutional image up-sampling | |
| CN107145578B (en) | Map construction method, device, equipment and system | |
| US9097542B2 (en) | Methods of pre-processing probe data | |
| US20180088228A1 (en) | Obstacle detection method and apparatus for vehicle-mounted radar system | |
| KR20190082071A (en) | Method, apparatus, and computer readable storage medium for updating electronic map | |
| WO2018181974A1 (en) | Determination device, determination method, and program | |
| CN111198560B (en) | Method and apparatus for predicting feature spatial attenuation using a variational self-encoder network | |
| US11927449B2 (en) | Using map-based constraints for determining vehicle state | |
| US11193789B2 (en) | Method, apparatus, and computer program product for identifying at-risk road infrastructure | |
| Zheng et al. | High definition map-based vehicle localization for highly automated driving: Geometric analysis | |
| US11549816B2 (en) | Systems and methods for controlling mapping information inaccuracies | |
| CN110926453A (en) | Obstacle positioning method and system | |
| US11327155B2 (en) | Radar sensor misalignment detection for a vehicle | |
| KR20220059573A (en) | System and method for updating high definition map | |
| WO2019188886A1 (en) | Terminal device, information processing method, and storage medium | |
| WO2019188820A1 (en) | Information transmission device, data structure, control method, program, and storage medium | |
| CN115235452A (en) | Intelligent parking positioning system and method based on UWB/IMU and visual information fusion | |
| CN116295508A (en) | Road side sensor calibration method, device and system based on high-precision map | |
| Welte et al. | HD map errors detection using smoothing and multiple drives | |
| US20210190502A1 (en) | Method and system for estimating the trajectory of an object on a map | |
| CN113566834A (en) | Positioning method, positioning device, vehicle, and storage medium | |
| Verentsov et al. | Bayesian framework for vehicle localization using crowdsourced data | |
| CN112987010A (en) | System and method for multi-radar mapping of robot | |
| CN111103584A (en) | Device and method for determining height information of objects in the surroundings of a vehicle | |
| CN114358038B (en) | Two-dimensional code coordinate calibration method and device based on vehicle high-precision positioning |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200327 |
|
| RJ01 | Rejection of invention patent application after publication |