Technical Field
The present invention relates to unmanned technical field, in particular to a based on the high-precision positioning and the lane recognition of high precision map automatic generation method.
Background Art
High-precision map as a unmanned field of scarce resources just need and, in the whole field of play a central role, can help the unmanned vehicle pre-sensing road complex information, such as slope, curvature, and course, combined with intelligent route planning, unmanned vehicle to make the correct decision-making, is unmanned vehicle-driving an indispensable data sources. Unmanned sensor needs to the collected information with the stored high-precision map contrast, determining the position and direction, in order to ensure the unmanned vehicle safe driving to a destination, so the accuracy of the high-precision map data acquisition for unmanned is very critical; the traditional high precision map that the manufacture of a large number of human indexed, not only wastes time, but also because of the marked manually generating the error rate of the error is also relatively high, and is not conducive to the development of unmanned, today's high-precision map making, because of the precision requirement is high, the computational process is complicated, the manufacturing process is relatively time consuming, provides a based on the high-precision positioning and on the basis of the lane recognition can be automatic splicing lane, time consuming shorter high precision map is very necessary.
Content of the invention
One of the purposes of the present invention is to solve the at least the above-mentioned problems, and to provide at least the rear and the advantages of the note.
The invention has another purpose is to provide a based on the high-precision positioning and the lane recognition of high precision map automatic generation method, in order to produce accurate to based on high precision positioning of the lane, and can automatically splicing the lane of high precision map, reduces the complexity of generating high precision map, and to avoid the traditional high-precision map making a large number of the human consumption and error rate is high.
In order to achieve these purposes according to the present invention and other advantages, provides a based on the high-precision positioning and the lane recognition of high precision map automatic generation method, comprising:
Step 1, by the time the alignment process, each time obtaining high-precision positioning data and lane line data synchronization, after synchronization to be of the high-precision positioning data of the position and the posture.
Step 2, by the steps 1 obtained in the lane after the synchronization of the data and high-precision positioning data to set up the map frame, and is stored in a database map frame.
Step 3, the new acquisition of the lane with the acquired data according to the data established lane all of said map frame matching, if the matching, using new acquisition of the lane data the establishment of a new map frame; if the match is successful, then in step 2 has been established in the map frame information updating, until the update is complete.
Step 4, through the step 2 and step 3 set up in the map frame among the frames in the frame does have some all smoothing processing, combined with the step 3 in the updated the map frame of the point set re-computation, in order to be treated through said interframe smooth the frame corresponding to the lane information representative of the three-time curve.
Step 5, through the three curve jointing, in order to produce the high precision map.
Preferably, the step 1 further includes:
In order to obtain the lane data time stamp for time alignment point, is aligned to the lane data time stamp of the high precision positioning data of the position p 'and attitude r' are:
P '=p + v (tm - Tl ); (1)
R '=r ω (tm - Tl ); (2)
Wherein tm The lane showing to obtain the data of the time stamp.
tl Said obtaining the high-precision positioning data of the time stamp.
P, r, v and ω are aligned before said the high-precision positioning data of the position, posture, the linear speed of the
Degrees and the angular velocity.
Preferably, step 2 further includes:
The map frame mainly comprising the following elements:
PF : The position of the frame space information; RF : The frame space information of the attitude.
CF : The lane information three-time curve; SF : Lane line sampling point set.
LF : The interframe topology information associated with the front and back of the frame.
The element are said to frame coordinate system, wherein the horizontal frame coordinate system for x shaft, longitudinal is y shaft, vertical x shaft, for z y axis direction of the shaft.
Preferably, step 2 further comprises:
The establishment of the map frame is based on the premise that represents the type of the lane and/or the color of the change and/or the lane line of the cutting and/or driving the unmanned vehicle lane and/or the map frame y axis direction length exceeds a threshold.
Preferably, step 3 matches in the conditions for success is:
The new acquisition of the lane with the acquired data according to the lane data established in all of said map frame one or a plurality of overlapped parts, and the part of the [...] reach to said threshold value; and/or
The newly acquired data with the lane has been established in the map frame of any one of the connected to the front.
Preferably, step 3 in the map frame for updating the information further includes:
Step C, through according to a certain interval to the lane represents the three curve sampling, sampling point set to be.
Step D, states the sampling point set by calculating the relative to the not more new the map frame of the position and the posture, and the position of the sampling point set of the projection to not more new the map frame of the in frame coordinate system, the coordinate system to be in the frame of the sampling point set.
Step E, if the frame in the coordinate system of the position of the sampling point set beyond the frame to limit the length of the coordinate system, by cutting out the excess, in order to as a new the lane data entering step C to the updating of the map frame.
Step F, through the not more new the map frame sampling point set and step D in the coordinate system of the frame in the merger of the sampling point set, in order to be combined according to a certain interval after re-sampling of the sampling point set, and the not more new the map frame sampling point set update for the re-sampling of the sampling point set.
Step G, through the step F re-sampling in the sampling point set to carry out three curve fitting, after fitting to be sampling point set, and the not more new the map frame sampling point set three times the results of curve fitting for the stated fitting after updating the sampling point set.
Step H, by calculating the new map frame has been built with of the connection relation of the map frame, connected to the previous completed meets the frame and frame after the correlation, in order to finish the established map frame of information updates.
Preferably, step 4 do in the interframe smooth processing is premised on the lane of the no new data input.
Preferably, step 4 frames in a smooth processing further includes:
The establishment of a three-time curve, wherein the curve of the point set [...] three times by the associated before stating, the frame and frame after the relations in the point before stating, the frame of the latter part of the precession of equinoxes collection and concentration of the first half of the [...] concentrated point integrates the mixing proceeds.
Through the latter half of the frame perpendicular to the point collection along the y shaft in the direction of the coordinate system, the projection to the three times on the curve, in order to be point set.
Through to the point before stating, the frame coordinate system with the point set in the latter half of the frame corresponding to the coordinate system of the projection point of the y shaft do after the smoothing treatment, in order to be point after smoothing, the thus smoothed frame; the subpoint [...] coordinate system integrates the smooth processing similar; point and point between smooth processing of the formula is:
P "=((1 - a) x + ax ', y, (1 - a) z + az') (3)
Wherein the front [...] smooth coefficient a=y/L; L represents the length of the frame.
The [...] smooth coefficient a=1 - y/L.
P=(x, y, z) before stating, the frame coordinate system representing the point on the latter part of any point of.
P '=(x', y, z ') representing the [...] coordinate system on the point of any point of the first half.
The invention comprises at least the following advantages:
The invention high-precision positioning data and data between the lane, and because of the high precision positioning data of the advance or lag, so that the existence of the time interval between the two, by the time the alignment process, the time aligned with the lane data acquisition time stamp, in order to guarantee data processing, the consistency of the data in time, in order to as high precision map generating the premise, data processing first need to carry on the high-precision positioning and lane line data between the coordinate system conversion; through the high-precision positioning data and the lane data is a priori conditions, the acquisition of the new lane data to set up the time of the map frame, the frame and the posture of the spatial information that the position of the facing, is the time after the alignment process of the high-precision positioning data of the position and the posture, the lane data changes, first needs to have been established with the map frame matching, demand overlap or with associated part, thus obtaining the lane data updated replacement, if matching is not successful, the obtained new lane data re-establishment of a new map frame, thus through matching updating and reconstruction, automatic completion of the splice of the lane; and then through the interframe smoothing processing, so that a smooth transition between frames, the overlapping portions of the [...] smoothly connected, in order to produce accurate to based on high precision positioning of the lane, and can automatically splicing the lane of high precision map, through the high-precision positioning and the lane of the a priori conditions obtaining, generating automatic splicing of the lane of the high precision map, reduces the complexity of generating high precision map, and to avoid the traditional high-precision map making a large number of the human consumption and high error rate, to unmanned reliable and safe running has important significance.
Other advantages of the present invention, objectives and features will reflect through the lower part of the note, segment will also be through to the study and practice of this invention is in the field of the technical understood.
Description of drawings
Figure 1 is the flow chart of the present invention based on the high-precision positioning and the lane recognition of high precision map automatic generation method;
Figure 2 is the schematic view of the present invention the frame coordinate system of the map frame comprising elements;
Figure 3 is the schematic view of the present invention can absorb almost the lane when the presence of the blind area;
Figure 4 is the schematic view of the invention the map frame information updating process can not only;
Figure 5 is the schematic view of the present invention can absorb almost the map frame after updating the information.
Mode of execution
The Figure below to the further detailed description of this invention, in order to make the technical personnel in the field specification can be on the basis of the implementation of the reference characters.
It should be understood, used herein such as "has", "comprising" and "including" terminology does not exclude one or a plurality of other components or a combination thereof the presence or added.
As shown in Figure 1, the present invention provides a based on the high-precision positioning and the lane recognition of high precision map automatic generation method, comprising:
Step 1, by the time the alignment process, each time obtaining high-precision positioning data and lane line data synchronization, after synchronization to be of the high-precision positioning data of the position and the posture.
Step 2, by the steps 1 obtained in the lane after the synchronization of the data and high-precision positioning data to set up the map frame, and is stored in a database map frame.
Step 3, the new acquisition of the lane with the acquired data according to the data established lane all of said map frame matching, if the matching, using new acquisition of the lane data the establishment of a new map frame; if the match is successful, then in step 2 has been established in the map frame information updating, until the update is complete.
Step 4, through the step 2 and step 3 set up in the map frame among the frames in the frame does have some all smoothing processing, combined with the step 3 in the updated the map frame of the point set re-computation, in order to be treated through said interframe smooth the frame corresponding to the lane information representative of the three-time curve.
Step 5, through the three curve jointing, in order to produce the high precision map.
In the above-mentioned scheme, high-precision positioning data and data between the lane, and because of the high precision positioning data of the advance or lag, so that the existence of the time interval between the two, by the time the alignment process, the time aligned with the lane data acquisition time stamp, in order to guarantee data processing, the consistency of the data in time, in order to as high precision map generating the premise, data processing first need to carry on the high-precision positioning and lane line data between the coordinate system conversion; through the high-precision positioning data and the lane for the a priori conditions data, the acquisition of the new lane data to set up the time of the map frame, the frame and the posture of the spatial information that the position of the facing, is the time after the alignment process of the high-precision positioning data of the position and the posture, the lane data changes, first needs to have been established with the map frame matching, demand overlap or with associated part, thus obtaining the lane data updated replacement, if matching is not successful, the obtained new lane data re-establishment of a new map frame, thus through matching updating and reconstruction, automatic completion of the splice of the lane; and then through the interframe smoothing processing, so that a smooth transition between frames, the overlapping portions of the [...] smoothly connected, in order to produce accurate to based on high precision positioning of the lane, and can automatically splicing the lane of high precision map, through the high-precision positioning and the lane of the a priori conditions obtaining, generating automatic splicing of the lane of the high precision map, reduces the complexity of generating high precision map, and to avoid the traditional high-precision map making a large number of the human consumption and high error rate, to unmanned reliable and safe running has important significance.
In one preferred embodiment, step 1 further includes:
In order to obtain the lane data time stamp for time alignment point, is aligned to the lane data time stamp of the high precision positioning data of the position p 'and attitude r' are:
P '=p + v (tm - Tl ); (1)
R '=r ω (tm - Tl ); (2)
Wherein tm The lane showing to obtain the data of the time stamp.
tl Said obtaining the high-precision positioning data of the time stamp.
P, r, v and ω are aligned before said the high-precision positioning data of the position, posture, the linear speed of the
Degrees and the angular velocity.
In the above-mentioned scheme in, through the lane data acquisition time setting the time stamp, to the lane data acquisition time stamp as a reference, the high-precision positioning the acquisition of data, because the data sampling point of time, there will be a certain advance or lag, to ensure that the lane data with high precision data in generating method in synchronous, through the formula 1 conversion, so as to obtain the algorithm of synchronized high-precision positioning data.
In one preferred embodiment, step 2 further includes:
The map frame mainly comprising the following elements:
PF : The position of the frame space information; RF : The frame space information of the attitude.
CF : The lane information three-time curve; SF : Lane line sampling point set.
LF : The interframe topology information associated with the front and back of the frame.
The element are said to frame coordinate system, wherein the horizontal frame coordinate system for x shaft, longitudinal is y shaft, vertical x shaft, for z y axis direction of the shaft.
In the above-mentioned scheme, as shown in Figure 2, frame coordinate system, comprising elements of the map frame.
In one preferred embodiment, step 2 further comprises:
The establishment of the map frame is based on the premise that represents the type of the lane and/or the color of the change and/or the lane line of the cutting and/or driving the unmanned vehicle lane and/or the map frame y axis direction length exceeds a threshold.
In the above-mentioned scheme, each of the representative of the map frame should be periodically synchronize to the lane data acquisition of a position and attitude, different position and attitude, namely different map frame based on identifying lane a high precision map, in the lane at the time of identification, the lane change of the data, in the premise of matching is not successful, will be the establishment of a new map frame, the lane data change, when the map frame is established, if the lane region in the map frame y axis direction exceeds a threshold, the position P of the frameF And attitude RF Are respectively XI as p 'and r', as shown in Figure 3, W the area shown in the lane data extraction to the map frame database, the lane data is obtained in the presence of the blind area y axis direction [0, yb ], The range in which the G in that, in the existing map frame in the region corresponding to the search data, if can be found, then the corresponding data to fill in to the blind area.
In one preferred embodiment, step 3 matches in the conditions for success is:
The new acquisition of the lane with the acquired data according to the lane data established in all of said map frame one or a plurality of overlapped parts, and the part of the [...] reach to said threshold value; and/or
The newly acquired data with the lane has been established in the map frame of any one of the connected to the front.
In the above-mentioned scheme, the lane of the newly obtained data first needs of matching the map frame has been established, in order to obtain the lane with the newly obtained data are partially overlapping or connected to the front of the map frame, thus confirm the newly obtained data of the position of the lane, so as to perform the updating of the map frame.
In one preferred embodiment, step 3 in the map frame for updating the information further includes:
Step C, through according to a certain interval to the lane represents the three curve sampling, sampling point set to be.
Step D, states the sampling point set by calculating the relative to the not more new the map frame of the position and the posture, and the position of the sampling point set of the projection to not more new the map frame of the in frame coordinate system, the coordinate system to be in the frame of the sampling point set.
Step E, if the frame in the coordinate system of the position of the sampling point set beyond the frame to limit the length of the coordinate system, by cutting out the excess, in order to as a new the lane data entering step C to the updating of the map frame.
Step F, through the not more new the map frame sampling point set and step D in the coordinate system of the frame in the merger of the sampling point set, in order to be combined according to a certain interval after re-sampling of the sampling point set, and the not more new the map frame sampling point set update for the re-sampling of the sampling point set.
Step G, through the step F re-sampling in the sampling point set to carry out three curve fitting, after fitting to be sampling point set, and the not more new the map frame sampling point set three times the results of curve fitting for the stated fitting after updating the sampling point set.
Step H, by calculating the new map frame has been built with of the connection relation of the map frame, connected to the previous completed meets the frame and frame after the correlation, in order to finish the established map frame of information updates.
In the above-mentioned scheme, as shown in Figure 4 and 5 shown, map frame information more new basic process was as follows:
According to a certain interval to the lane of the representatives of the three curve A sampling, sampling point set by SA ;
Calculating A relative to the current map frame F position posture, and the sampling point set SA The position posture projection to on the F, write ![]()
If ![]()
Beyond the current map frame F length limited, will exceed the part of the cutting out and its as a new lane data re-map frame update step;
The current map frame F S sampling pointF with the ![]()
The merger, and according to a certain interval re-sampled to obtain SF ', And will SF Updating is updated to the SF ';
To SF ' Do three curve fitting to obtain CF ', And CF C is updated toF ';
For a new map frame calculating its with the connection relation of the map frame, updated accordingly is connected with the front and back of the meets the frame L [...] relationsF .
In one preferred embodiment, step 4 do in the interframe smooth processing is premised on the lane of the no new data input.
In the above-mentioned scheme, map frame information has been built on the premise that the update is complete, the absence of new lane data line is obtained, otherwise once the new lane data is obtained, the new lane data will enter with the matching of the map frame has been built, or the establishment of a new map frame or enter the updating of the map frame in construction, so only in the absence of new lane data input under the condition of, in order to ensure the completion of the map frame has been updated to carry out interframe smooth processing.
In one preferred embodiment, step 4 frames in a smooth processing further includes:
The establishment of a three-time curve, wherein the curve of the point set [...] three times by the associated before stating, the frame and frame after the relations in the point before stating, the frame of the latter part of the precession of equinoxes collection and concentration of the first half of the [...] concentrated point integrates the mixing proceeds.
Through the latter half of the frame perpendicular to the point collection along the y shaft in the direction of the coordinate system, the projection to the three times on the curve, in order to be point set.
Through to the point before stating, the frame coordinate system with the point set in the latter half of the frame corresponding to the coordinate system of the projection point of the y shaft do after the smoothing treatment, in order to be point after smoothing, the thus smoothed frame; the subpoint [...] coordinate system integrates the smooth processing similar; point and point between smooth processing of the formula is:
P "=((1 - a) x + ax ', y, (1 - a) z + az') (3)
Wherein the front [...] smooth coefficient a=y/L; L represents the length of the frame.
The [...] smooth coefficient a=1 - y/L.
P=(x, y, z) before stating, the frame coordinate system representing the point on the latter part of any point of.
P '=(x', y, z ') representing the [...] coordinate system on the point of any point of the first half.
In the above-mentioned scheme, is related to the relationship of the front and back respectively of the two frame F1 And F2 , Taking the front [...]![]()
The latter part of the point set ![]()
And [...]![]()
The first half of the point set ![]()
Mixed, fitting three curve Cm ; For the point set ![]()
Y will be along the direction perpendicular to the axis of the projection to the three curve Cm The upper, get the point set ![]()
For the frame F1 Coordinate system point set ![]()
In each of a point P=(x, y, z) and instead of ![]()
The corresponding point P '=(x', y, z '), according to the following formula after smoothing of the point P "=((1 - a) x + ax', y, (1 - a) z + az '), wherein the smoothing coefficient a=y/L, L for the frame F1 The length of the.
Frame F2 Point set of smooth similar, the difference between the smooth coefficient into a=1 - x/L; in particular, for the point set![]()
Y will be along the direction perpendicular to the axis of the projection to the three curve Cm The upper, get the point set ![]()
For the frame F2 Coordinate system point set ![]()
In each of a point P=(x, y, z) and instead of ![]()
The corresponding point P '=(x', y, z '), according to the following formula after smoothing of the point P "=((1 - a) x + ax', y, (1 - a) z + az '), wherein the smoothing coefficient a=1 - y/L, L frame F2 The length of the;
For the finish point set smooth frame, according to the updated point set re-calculation of the frame three curve C laneF .
Although the embodiments of the invention have been disclosed above, but not limited to the specification and embodiments set out in the application, it is fully can be suitable for various suitable for the field of this invention, familiar in the case of the field personnel, can be easily achieved in addition changes, therefore without departing from the claims and the equivalent of the scope of the general concept of defined, the invention is not limited to the specific details shown and described here of the legend.
Automatic generation method for high-precision map based on high precision positioning and lane line recognition
400-4929-0909
contact@catarc.com.cn
