CN107044853A - Method and apparatus for definitely calibration method and device and for positioning - Google Patents

Method and apparatus for definitely calibration method and device and for positioning Download PDF

Info

Publication number
CN107044853A
CN107044853A CN201610941150.1A CN201610941150A CN107044853A CN 107044853 A CN107044853 A CN 107044853A CN 201610941150 A CN201610941150 A CN 201610941150A CN 107044853 A CN107044853 A CN 107044853A
Authority
CN
China
Prior art keywords
terrestrial reference
data
constellation
positioning
terrestrial
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.)
Granted
Application number
CN201610941150.1A
Other languages
Chinese (zh)
Other versions
CN107044853B (en
Inventor
W·尼姆
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Robert Bosch GmbH
Original Assignee
Robert Bosch GmbH
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Robert Bosch GmbH filed Critical Robert Bosch GmbH
Publication of CN107044853A publication Critical patent/CN107044853A/en
Application granted granted Critical
Publication of CN107044853B publication Critical patent/CN107044853B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/04Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by terrestrial means
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C23/00Combined instruments indicating more than one navigational value, e.g. for aircraft; Combined measuring devices for measuring two or more variables of movement, e.g. distance, speed or acceleration

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Navigation (AREA)

Abstract

The present invention relates to a kind of for the method for the terrestrial reference (L1, L2) for determining to be positioned for the object to motion.Wherein, this method has the step of at least one constellation data record (150) is generated in the case of using regular (140) and landmark data (130) is determined.Here, the landmark data represents the terrestrial reference in the surrounding environment of object determined by sensor.Determining rule can be employed for determining at least one pair of terrestrial reference suitable for positioning by landmark data, wherein, between the track (105) of object and described pair of the first terrestrial reference (L1) ratio between horizontal spacing (a) and the terrestrial reference spacing (L) between described pair of the first terrestrial reference (L1) and the second terrestrial reference (L2) is in the range of predetermined fit value.Here, at least one constellation data record is with position data of a pair of terrestrial references relative to the geometry constellation of object suitable for positioning.

Description

Method and apparatus for definitely calibration method and device and for positioning
Technical field
The present invention relates to the apparatus and method described in a kind of content according to independent claims.Present disclosure is also Computer program.
Background technology
Some systems for being used for highly automated traveling, robot and augmented reality can be for example based on to high-acruracy survey Terrestrial reference analysis.
The content of the invention
Under the background, following methods are proposed with scheme described herein, proposed in addition a kind of using in methods described At least one method device and finally propose a kind of corresponding computer program according to independent claims. The favourable extension of the device provided in the independent claim can be realized by the measure implemented in the dependent claims Scheme and improvement project.
According to the embodiment of the present invention, the terrestrial reference that can be especially to provide optimization for positioning is obtained, for example based on Pre-existing mark and the geometrical relationship of self-position provide the terrestrial reference and obtained.Can especially it be to provide to for height The automatically optimal acquisition of the terrestrial reference for the positioning that traveling, robot and augmented reality are carried out.Here, positioning can be especially by Realized using the terrestrial reference of the only only a part quantity for the positioning that is particularly suitable for use in, the constellation of the terrestrial reference can make position error most It is small.
Advantageously, it can particularly realize according to the embodiment of the present invention and optimally select and significantly reduce terrestrial reference Quantity, the terrestrial reference can be used for for example supermatic vehicle, robot are positioned.Recognized by means of selection course For be particularly suitable for use in positioning terrestrial reference can reduce the data to be transmitted quantity in the case of realize more reliably, it is rapider Ground, less consuming storage, less expend calculate ground and more accurately positioned.
A kind of ground calibration method for being used to determine to be used to position the object of motion, wherein, this method has following Step:
At least one constellation data is generated in the case of using rule and landmark data is determined to record, wherein, terrestrial reference number According to the terrestrial reference in the surrounding environment of object that is obtained by sensor of expression, wherein, determining rule can be used, so as to from ground At least one pair of terrestrial reference that data determination is applied to positioning is marked, wherein, between the track of object and described pair of the first terrestrial reference Ratio between horizontal spacing and terrestrial reference spacing between described pair of the first terrestrial reference and the second terrestrial reference is in predetermined be adapted to It is worth in scope, wherein, at least one constellation data record is with the geometry constellation for being applied to position a pair of terrestrial references relative to object Position data.
This method can for example exist for example in software or hardware or in the combining form being made up of software and hardware Performed in control unit.Object may, for example, be vehicle, particularly motor vehicle, for example designed for automatic running at least in part Vehicle, robot or similar device.The surrounding environment of object can be geographic area, and object can be arranged in the geographic region In domain.Terrestrial reference in the surrounding environment of object can be determined for example by means of at least one sensor device.For described true Fixed method for example can be implemented with reference to designated vehicle, and the designated vehicle is designed for determining in the surrounding environment of object Terrestrial reference.Designated vehicle can have at least one sensor device.Thus at least one constellation data record, which can have, is adapted to Terrestrial reference position data, wherein, position data can be related to predetermined coordinate system, such as numerical map.Terrestrial reference can have Position data, view data or the class likelihood data extracted from sensing data.Fit value scope can include a pair of terrestrial references Maximum fit value, wherein, maximum fit value can correspond to use determine rule in the case of determine in horizontal spacing and The ratio value of ratio between terrestrial reference spacing.
According to one embodiment, methods described can have from original terrestrial reference extracting data landmark data the step of, institute State landmark data and find value again with unrelated with visibility conditions, the value that finds again finds value again more than minimum, The original terrestrial reference tables of data shows the total amount of terrestrial reference in the surrounding environment of object.Here, can be using being extracted Landmark data in the case of the step of implement the generation.Again the find value unrelated with visibility conditions can be in use pair As being generated in the case of the sight that exists between terrestrial reference under different visibility conditions.The advantage that the embodiment is provided It is that following terrestrial references are only determined for one-time positioning, the terrestrial reference can reliably be known again in adverse conditions in itself Not.It is possible thereby to improve the reliability of positioning and improve the high speed of positioning if necessary.
In addition, methods described can have in the case of using rule is determined and additionally or alternatively consider ground The step of marking predetermined fit value scope in the case of data, the position of object, the speed of the track of object and/or object.It is described Embodiment is provided the advantage that, can also be used under the different motion conditions of object for the suitable suitable positioning of terrestrial reference Rule.
Methods described, which can also have, to be applied to positioning extremely from least one constellation data record generated The step of position data of few a pair of terrestrial references is added to the map datum of numerical map.Here, numerical map can represent reference Coordinate system.It is used to position in addition, numerical map could be designed to be used by the object moved.
It is also proposed that a kind of method for being used to position the object of motion, wherein, this method has following step:
At least one constellation data record is read, the constellation data record represents an embodiment party according to preceding method The data record of formula generation;
Know in the case of the position data for the terrestrial reference for being applied to positioning for considering to record from least one constellation data Terrestrial reference not in the image of the surrounding environment of object;And
Implement positioning in the case of the position data of the terrestrial reference recognized in the step of using in the identification.
This method can for example exist for example in software or hardware or in the combining form being made up of software and hardware Performed in control unit.Here, the method for positioning can combine the object of motion to implement.Method for positioning also can Enough combine is previously described for the embodiment of definitely calibration method to implement.
According to an embodiment, the step of the implementation in, using landmark data, the position of object, object Implement the positioning in the case of the speed of track and/or object.The embodiment is provided the advantage that, according to the fortune of object Emotionally condition is realized reliable and accurately positioned.
The step of this method can also have the image for the surrounding environment for obtaining object.Here, at least one can used The step of implementing the measure in the case of individual acquisition device.Here, the object has at least one acquisition device.The reality The mode of applying is provided the advantage that, in order to be positioned to object, can use corresponding to accurately and currently for surrounding environment Move the image of constellation.
In addition scheme described herein proposes a following device, and it is designed for performing in corresponding equipment, controlled The step of making or realize method described herein.Can be rapidly in the form of device also by the implementation variant of the present invention And efficiently realize the purpose where the present invention.
Here, device can be understood as electronic installation, the electronics process sensor signal and passed according to described Sensor signal output control signal and/or data-signal.Described device can have interface, and the interface can be with the side of hardware Formula and/or constituted in the way of software.In the design of hardware, interface may, for example, be the portion of so-called ASIC systems Part, the part includes the different functions of described device.However it is also possible that, the interface be itself integrated circuit or Person is made up of discrete element at least in part.In the design of software, the interface is software module, the software mould Block is for example present in micro controller beside other software module.
Also advantageously a kind of computer program product or the computer program with program coding, the computer program Machine readable carrier or storage medium, such as semiconductor memory, harddisk memory or optical memory can be stored in, and It is used to implement, realize and/or control according to aforementioned embodiments when program product or program are performed on computer or device Any one of method the step of.
Brief description of the drawings
Embodiments of the invention are shown in the drawings and are set forth in the following description.In accompanying drawing:
Fig. 1 shows the schematic illustration of the determining device according to one embodiment;
Fig. 2 shows the schematic illustration of the positioner according to one embodiment;
Fig. 3 shows the flow chart of the method for the determination according to one embodiment;
Fig. 4 shows the flow chart of the method for being used to position according to one embodiment;
Fig. 5 shows the schematic illustration for the determining device 110 for coming from Fig. 1;
Fig. 6 shows the schematic illustration for being used to record the camera model of terrestrial reference according to one embodiment;
Fig. 7 shows the terrestrial reference L and the schematic illustration of position error according to one embodiment;
Fig. 8 shows the schematical top view on error model according to one embodiment;
Fig. 9 shows the schematic illustration of the constellation of the terrestrial reference according to one embodiment;
Figure 10 shows the error chart of angle according to the observation;
Figure 11 shows to be used for the schematic illustration of the geometric error model of the first constellation of road sign according to one embodiment;
Figure 12 shows the schematic illustration for the geometric error model of the second constellation of road sign according to one embodiment,
Figure 13 shows the error chart according to one embodiment angle φ according to the observation;
Figure 14 shows the schematic illustration of the evaluated error according to one embodiment;
Figure 15 shows the schematic illustration for being used to determine terrestrial reference viewing angle according to one embodiment;
Figure 16 shows the optimization chart on terrestrial reference constellation according to one embodiment;
Figure 17 shows the optimization chart on terrestrial reference constellation according to one embodiment;With
Figure 18 shows the optimization chart on terrestrial reference constellation according to one embodiment.
In the description below of the advantageous embodiment of the present invention, for shown in different figures and act on similar Element use same or similar reference, wherein, cancel repeat specification to the element.
Embodiment
Fig. 1 shows the schematic illustration of the determining device 110 according to one embodiment.Where it determines that device 110 only shows It is arranged in measurement vehicle 100 to example property.Embodiment according to Fig. 1, measures the edge on a highway herein of vehicle 100 Track 105 to travel.In order to better illustrate, measurement vehicle 100 is exaggerated ground and schematically shown again in Fig. 1.
First terrestrial reference L1 with the horizontal spacing a of track 105 arrange.Second terrestrial reference L2 is between the terrestrial reference away from the first terrestrial reference L1 Away from L arrangements.Embodiment according to Fig. 1, terrestrial reference spacing L extends along track 105 herein.Therefore, the second terrestrial reference L2 has There is the horizontal spacing a with track 105.For purposes of illustration, also show that in Fig. 1 terrestrial reference L1 and L2 with measurement vehicle 100 it Between sight.
Measuring vehicle 100 has determining device 110.Determining device 110 is entered designed for determining to be used for the object of motion The terrestrial reference of row positioning.Therefore, determining device 110 has generating means 120.In other words, in Fig. 1 from determining device 110 Generating means 120 are only shown.Generating means 120 are designed in the case where using landmark data 130 and measure rule 140 Generate at least one constellation data record 150.
The object of motion is, for example, an other vehicle, such as user vehicle.Other vehicle or user's vehicle can be with Moved at least partially along the same or similar track in track 105 with measuring vehicle 100.This is in fig. 2 also in detail Explanation.
Here, landmark data 130 represents the terrestrial reference in surrounding environment obtained by sensor, is herein the first terrestrial reference L1 With the second terrestrial reference L2, the object of motion can be moved in the surrounding environment.Landmark data 130 is for example with the first terrestrial reference L1 With the second terrestrial reference L2 position data.
Here, it is noted that the embodiment according to Fig. 1 is exemplarily illustrated only two terrestrial references L1 and L2.Root There can also be other terrestrial reference according to a further embodiment.
Measure rule 140 can be determined device 110, particularly generating means 120 and use, to cause by landmark data 130 determine at least one pair of terrestrial reference suitable for positioning, wherein, the horizontal stroke between the track of object and described couple of the first terrestrial reference L1 The ratio between terrestrial reference spacing L to spacing a and between described couple of the first terrestrial reference L1 and the second terrestrial reference L2 is predetermined In the range of fit value.
At least one can be had by means of the constellation data record 150 of determining device 110, the particularly generation of generating means 120 There is position datas of a pair of terrestrial reference L1 and L2 relative to the geometry constellation of the object suitable for positioning.
Using in the case of determining device 110, it is possible to achieve optimally determined in the numerical map of high precision or Person surveys and draws the terrestrial reference for example for highly automated traveling and robot and data receiver.Measure vehicle 100 or determine car 100 particularly have RTK-GPS devices (Real Time Kinematic Global Positioning System, in real time Dynamic global positioning system) and numerical map, thus vehicle-mounted be used to determine the sensor of vehicle-periphery, such as video dress Put, radar, laser radar etc..In the first step, from the data of the vehicle-mounted sensing device, particularly from landmark data The terrestrial reference L1 and L2 or following characteristics are extracted in 130, the feature is particularly suitable for use in also in other visibility and illuminance It is lower to be picked up.Example is the feature of guideboard, signal lamp, mark on road surface etc..These typically also substantial amounts of features exist Use in the case of determining device 110 by this spline filter, reduction or rarefaction, the geometry constellation for causing to realize high position precision Feature is retained.Favourable terrestrial reference constellation is to cause the sight squarely phase as far as possible between terrestrial reference and measurement vehicle 100 herein The constellation of friendship, this causes the position error of minimum again.The feature then using its characterize symbol and corresponding 3D coordinates as Constellation data record 150 is write in positioning map and is then used as the terrestrial reference L1 and L2 for positioning.
Fig. 2 shows the schematic illustration of the positioner 210 according to one embodiment.Here, positioner 210 is arranged in In the object 200 of motion, the object of the motion is designed as vehicle 200 or user's vehicle 200 merely illustratively.Except user Beyond vehicle 200, the diagram in Fig. 2 is corresponding to the diagram for coming from Fig. 1.Therefore, user's vehicle 200 on highway along track 105 travelings.For the purpose better illustrated, user's vehicle 200 is also additionally exaggerated ground and schematically shown in fig. 2 Go out.
User's vehicle 200 has positioner 210.Positioner 210 is designed for object 200 or user's car to motion 200 are positioned.Here, positioner 210 is designed for right in the case where using at least one constellation data record 150 User's vehicle 200 is positioned, and the constellation data record is filled by means of the determining device 110 or similar determination for coming from Fig. 1 Put generation.In other words, positioner 210 is designed for recording 150 or class using at least one constellation data for coming from Fig. 1 As the object 200 or user's vehicle 200 of motion are positioned in the case of constellation data record.
Therefore, positioner 210 has reading device 212, identifying device 214 and performs device 216.Here, reading dress 212 are put designed for reading at least one constellation data record 150.Identifying device 214 is designed for considering to come from least one In the case of the terrestrial reference L1 and L2 that are applied to position of individual constellation data record 150 position data around user's vehicle 200 Terrestrial reference is recognized in the image of environment.Here, representing the image of surrounding environment by view data 220.View data 220 for example exists It is measured in the case of sensor device using user's vehicle 200.Here, identifying device 214 is designed for receiving or recalling View data 220.Thus, identifying device 214 is designed for recording using at least one constellation data in view data 220 Identification is suitable for the terrestrial reference of positioning, herein such as terrestrial reference L1 and L2 in the case of 150.Performs device 216 is designed for using Positioned in the case of the terrestrial reference L1 and L2 that are recognized by means of identifying device 214 position data.
Thus realized and optimally selected in running operation, particularly for highly automated by means of positioner 210 The terrestrial reference for being used to position of traveling and robot.In order to which automatic running function uses terrestrial reference in vehicle 200 or user's vehicle 200 During L1 and L2, for example set for user's vehicle 200 and read the GPS, digitally that 150 are recorded with least one constellation data Figure, onboard sensor and positioning map.Particularly continuously attempted in automatic running at least in part, for user's vehicle 200 current location, speed and position is by means of positioner 210 and is for example found again by means of onboard sensor device It is stored in terrestrial reference, such as terrestrial reference L1 and L2 in positioning map.Here, positioner 210 is designed for identifying or looking for To in the case of terrestrial reference exactly, i.e. it is then preferred that terrestrial reference L1 and L2, are marked on current location and the feelings of user's vehicle 200 describedly The terrestrial reference geometry constellation for causing positioning precision maximum under condition.
Fig. 3 shows the flow chart of the method 300 for the determination according to one embodiment.Methods described 300 can Implement the terrestrial reference for determining to be used to position the object of motion.Here, the method 300 for the determination can be combined Come from Fig. 1 determining device or similar determining device to implement.
Methods described 300 has generates at least one constellation data note in the case of using rule and landmark data is determined The step 310 of record 150.Here, landmark data represents the terrestrial reference in the surrounding environment of object determined by sensor, wherein, Rule is determined to be used, to be determined by landmark data to be applied at least one pair of terrestrial reference of positioning, wherein, in object Horizontal spacing between track and described pair of the first terrestrial reference and the terrestrial reference between described pair of the first terrestrial reference and the second terrestrial reference Ratio between spacing is in the range of predetermined fit value.Here, at least one constellation data record, which has, is applied to positioning Position data of a pair of terrestrial references relative to the geometry constellation of the object.
The step of methods described 300 has the step 320, the predetermined step 330 and/or the addition of the extraction 340.Here, the step 320 of the extraction and the predetermined step 330 can be before the steps 310 of the generation by reality Apply, wherein, the step 340 of the addition can be carried out after step 310.Here, in the step 320 of the extraction from Original terrestrial reference extracting data landmark data, the landmark data have it is unrelated with visibility conditions find value again, it is described Value is found again and finds value again more than minimum, and the original terrestrial reference tables of data shows total amount of in the surrounding environment of object Terrestrial reference.Here, implementing the step 310 of the generation in the case where using the landmark data being extracted.Described predetermined In step 330, using determine rule in the case of and/or consider landmark data, the position of object, the track of object and/ Or make a reservation for the fit value scope in the case of the speed of object.In the step 340 of the addition, generated by least one Constellation data record will suitable for the position data of at least one pair of terrestrial reference of positioning be added to numerical map map datum.
Fig. 4 shows the flow chart of the method 400 for being used to position according to one embodiment.Methods described 400 can be implemented For the object implementatio8 positioning to motion.Here, the method 400 for the positioning can combine the positioner for coming from Fig. 2 Or similar positioner and additionally combine come from Fig. 1 determining device or similar determining device to implement if necessary.
Methods described 400 has the step 410 for reading at least one constellation data record.Here, constellation data records root Generated according to Fig. 3 method or similar method.The method or similar method that methods described 400 can combine Fig. 3 are implemented.
The step 420 for the identification that can then implement in the step 410 on the reading in methods described 400 In, in object in the case of the position data for the terrestrial reference for being applied to positioning for considering to record from least one constellation data Terrestrial reference is recognized in the image of surrounding environment.In methods described 400 in the step 430 of the implementation of a continuation, make Implement the positioning in the case of the position data of the terrestrial reference recognized in the step of used in the identification.
According to one embodiment, in the step 430 of the implementation, landmark data, the position of object, object are being used Implement the positioning in the case of the speed of track and/or object.
According to a further embodiment, the method 400 for positioning also has the image of the surrounding environment of measure object Step 440.Implement here, the step 440 of the measure can be particularly before the step 410 of the reading.Alternatively, The step 440 of the measure can also be implemented before the step 420 of the identification.
Fig. 5 shows the schematic illustration for the determining device 110 for coming from Fig. 1.Here, determining device for purposes of illustration Dividually shown with measurement vehicle 100.Here, measurement vehicle 100 is moved along track 105 or pre-existing section 105. This, measurement vehicle 100 have sensing device or at least one be used for determine surrounding environment sensor device, such as video dress Put, laser radar or similar device and the sensor device for determining self-position, and the measurement vehicle sets There is the electronics visual field.
Determining device 110 designed for receive or read for example measurement vehicle 100 visual field sensor (video-unit, swash Optical radar) view data 532, the section data 534 on pre-existing section and the electronics visual field, on for assessing The vehicle position data 536 (RTK-GPS) and calculating data 542 of the model of position error.
In Fig. 5 diagram, three processing square frames 522,524 and 526 are merely illustrative from determining device 110. In the first processing square frame 522, in the case where using section data 534 and vehicle position data 536 and calculating data 542 Calculate favourable terrestrial reference constellation.Can be by so-called (the ROI=Region of of ROI data 538 by the first processing square frame 522 Interests, area-of-interest) it is sent on second processing square frame 524.Realized in second processing square frame 524 in ROI data Terrestrial reference is found in 538.This can use SIFT (Scale-invariant feature transform) namely yardsticks not Become Feature Conversion and/or DIRD- features (DIRD=Illumination Robust Descriptor illumination robusts describer). Terrestrial reference is found in addition, being realized in two processing square frames 524 in the case where using view data 532.By second processing square frame 524 There is iteration loop and return to the first processing square frame 522.In the 3rd processing square frame 526, using coming from second processing square frame 524 Terrestrial reference and calculate data 542 in the case of realize selection can with high position precision terrestrial reference constellation.
Determining device 110 is designed for output or provides map datum 550, wherein, map datum 550 represents have at least The electronic map 555 or polar plot of the numerical map of one constellation data record.Therefore, electronic map 555 represents multiple terrestrial references LM and track 105 or section 105.In other words, electronic map 555 represents the road of the terrestrial reference constellation with the optimization measured Section 105.
In other words, Fig. 5 is shown for optimally determining the device that terrestrial reference or terrestrial reference are determined or drawn.Measure vehicle 100 Ambient sensors (such as laser radar, video camera), RTK-GPS and numerical map e.g. equipped with output image or The electronics visual field.In at least one geometrical property, such as curvature, turning radius, the gradient or the class using pre-existing section 105 As characteristic and measurement vehicle 100 current location in the case of determine that what is that have in the case where using determining device 110 The terrestrial reference constellation of profit, suitable terrestrial reference is then found for this terrestrial reference constellation in the ambient sensors of output image, Wherein, it may be iterated, and the 3D positions of the terrestrial reference are then logged into numerical map 555.
Fig. 6 shows the schematic illustration for being used to record the camera model 600 of terrestrial reference according to one embodiment.Video camera Model 600 also illustrates that the selection of sensor coordinates system and the sensor model for recording terrestrial reference.Camera model 600 can With reference to the determining device for coming from Fig. 1 and come from Fig. 3 for the determination method and/or come from Fig. 2 positioner with And come from Fig. 4 method for the positioning and used.
Camera model 600 by using based on polar coordinate system/cylindrical coordinate, wherein, origin is arranged in Jiao of video camera In point F and realize and be imaged on cylindrical surface either imaging surface 610 or image surface 610, as similar to panorama As video camera.Here, regarding between the object-point LM and focus F of the point LM or terrestrial reference in real world are marked to Fig. 6 Line LOS.Sight LOS and imaging surface 610 intersection point represent picture point IL.Focus F is surrounded by imaging surface 610.
Imaging surface 610 is for example allocated to every 360 degree of RcamThe video camera of the image resolution ratio of pixel.Line of reference 615 Imaging surface 610 is extended to from focus F.In addition, figure 6 illustrates viewing angle φ, the viewing angle is in sight LOS and ginseng Deploy according between line 615.
Fig. 7 shows the schematic illustration of the terrestrial reference L and position error E according to one embodiment.Fig. 7 is particularly illustratively Demarcate the model of position error E.Terrestrial reference position error E is expressed as ball or circle with diameter E, wherein, terrestrial reference LM actual bit Put not with described circular or described ball.
Fig. 8 shows the schematical top view on error model according to one embodiment.In other words, Fig. 8 is bowed with 2D View shows the error model of geometry.Here, error model is for example based on the camera model for coming from Fig. 6.
The imaging surface 610 for illustrating cylindrical video camera is overlooked with 2D in fig. 8.In addition, exemplarily giving in fig. 8 Feeding two terrestrial references L1 and L2 gone out in coordinate system WKOS, it is observed that described under picture point IL1 and IL2 from cameras Terrestrial reference.Also two sights LOS1 and LOS2 are marked, the sight extends to the focus of video camera from terrestrial reference respectively, and the focus is Desired position Pf.Thus, the first sight LOS1 extends to desired position Pf from the first terrestrial reference L1 through the first picture point IL1, Wherein, the second sight LOS2 extends to desired position Pf from the second terrestrial reference L2 through the second picture point IL2.Sight LOS1 and LOS2 Enclose at an angle or viewing angle φ.
In view of for reference to the determining device for coming from Fig. 1 and come from Fig. 3 for the determination method and/or go out From Fig. 2 positioner and come from Fig. 4 scheme positioned for the method for the positioning, in the terrestrial reference L1 provided and Corresponding sight LOS1 and LOS2 is set up in L2 and the picture point IL1 and IL2 that are determined coordinate (WKOS), the sight is being expected Position Pf in intersect.Thus the constellation or geometry constellation of positioning can be used for by one being generated between terrestrial reference L1 and L2.
Fig. 9 shows the schematic illustration of the constellation of the terrestrial reference L1 and L2 according to one embodiment.Terrestrial reference L1 and L2 constellation For example corresponding to the constellation for coming from Fig. 8.Here, terrestrial reference L1 and L2 constellation are shown on desired position Pf, it is described desired Object, such as vehicle or the similar object of positional representation motion.Here, the unknown position Pf of vehicle can be by two by car The known terrestrial reference L1 and L2 that (panorama) video camera is observed are set up.
The main source of error for definitely target constellation is exemplarily illustrated in Figure 10 to 14.Here, on the one hand relating to And terrestrial reference L1 and L2 site error, and the evaluated error for the picture point being on the other hand related in camera plane.
Figure 10 shows the error chart according to one embodiment angle φ according to the observation.Here, the observation in units of degree Angle φ represented on axis of abscissas, wherein, by worst error DmaxWith the ratio of the terrestrial reference position error E formation shown in Fig. 7 Represent on the vertical scale.The curve map 1010 of the error distribution on viewing angle φ is also showed that herein.Curve map 1010 is big There is minimum value under about 90 degree of viewing angle φ.
Figure 11 shows to be used for the schematic of the geometric error model of road sign L1 and L2 the first constellation according to one embodiment Diagram.It is shown in which the first road sign L1 and the second road sign L2, institute's road sign is stated with its position error E and intersecting accordingly regarded Line and the viewing angle φ with the first value.Embodiment according to Fig. 1, viewing angle φ is acute angle.Sight it is intersecting Region represents to come from Figure 10 worst error D herein in the case where considering corresponding position error Emax.It is desirable here that position Pf is put to be arranged in intersecting area.
Figure 12 shows to be used for the schematic of the geometric error model of terrestrial reference L1 and L2 the second constellation according to one embodiment Diagram.Here, Figure 12 diagram is corresponding to the diagram for coming from Figure 11, in addition to following situations, viewing angle φ is according to Figure 12 Shown embodiment is obtuse angle, represents worst error DmaxIntersecting area be less than Figure 11 in intersecting area.
Figure 13 shows the error chart according to one embodiment angle φ according to the observation.Here, Figure 13 error chart phase Ying Yu comes from Figure 10 error chart, in addition to following situations, and the maximum for such as 0.1 meter of terrestrial reference position error E is missed Poor DmaxRepresent on axis of ordinates.The curve map 1310 of the error distribution on viewing angle φ is also showed that herein.Curve map 1310 have minimum value under about 90 degree of viewing angle φ.
With reference to figures 10 to 13 supplementary notes, road sign L1 and L2 terrestrial reference position error or site error E cause rhombus Face or intersection, the desired position Pf of the object of motion is arranged in the face of the rhombus or intersection.Worst error Dmax It is relevant with viewing angle φ and terrestrial reference position error E, as also as being evident that following formula:
Dmax/ E=2cos (φ/2)/tan (φ/2), 0 < φ≤90 °
Dmax/ E=2cos ((180 ° of-φ)/2)/tan ((180 ° of-φ)/2), 90 < φ≤180 °
Dmax=2E cos (φ/2)/tan (φ/2), 0 < φ≤90 °
Dmax=2E cos ((180 ° of-φ)/2)/tan ((180 ° of-φ)/2), 90 < φ≤180 °
Figure 14 shows the schematic illustration of the evaluated error U according to one embodiment.In other words, Figure 14 is shown in shooting The evaluated error U of machine plane or the picture point in imaging surface 610.Figure 14 illustrates terrestrial reference LM, it is marked with describedly and video camera Focus F spacing R arrangement.Here it is shown that in dimimagerResolution ratio image in Sub-pixel estimation precision Δi.Son Pixel measurement accuracy ΔiCause the site error or evaluated error U of the terrestrial reference LM determined in spacing R.Here, drawing survey Determining error is:
U=(2 π R Δsi)/dimimager
In the Sub-pixel estimation precision Δ using 0.5 pixeliResolution ratio dim is determined with the horizontal image of 2000 pixelsimager In the case of, for example it can show that evaluated error U error amount is 0.02 meter for 10 meters of spacing, for 20 meters of spacing R The error amount for drawing evaluated error U is 0.03 meter, show that evaluated error U error amount is 0.05 meter for 30 meters of spacing R, right The error amount that evaluated error U is drawn in 40 meters of spacing R is 0.06 meter, and evaluated error U error is drawn for 50 meters of spacing R It is worth for 0.08 meter, show that evaluated error U error amount is 0.09 meter for 60 meters of spacing R, and obtain for 70 meters of spacing R The error amount for going out evaluated error U is 0.11 meter.
Figure 15 shows the schematic illustration for being used to determine terrestrial reference viewing angle φ according to one embodiment.Figure in Figure 15 Show the partial sector of diagram herein similar to Fig. 1 or Fig. 2.Here, figure 15 illustrates the object of the motion in vehicle form, The vehicle is, for example, the measurement vehicle for coming from Fig. 1 or the user's vehicle for coming from Fig. 2.Embodiment according to Figure 15, The vehicle is travelled on highway along track 105 herein.In addition, showing fortune of the vehicle along track 105 with kinematic parameter x It is dynamic, wherein, kinematic parameter x represents vehicle location, speed or similar parameter.
First terrestrial reference L1 with the horizontal spacing a of track 105 arrange.Second terrestrial reference L2 is between the terrestrial reference away from the first terrestrial reference L Away from L arrangements.Embodiment according to Figure 15, terrestrial reference spacing L extends along track 105 herein.Therefore, the second terrestrial reference L2 is same Sample has the horizontal spacing a with track 105.In addition, figure 15 illustrates the sight LOS1 between terrestrial reference L1 and L2 and vehicle and LOS2.Here, the first sight LOS1 extends to vehicle from the first terrestrial reference L1, wherein, the second sight LOS2 prolongs from the second terrestrial reference L2 Reach vehicle.Terrestrial reference viewing angle φ deploys between the first sight LOS1 and the second sight LOS2.
As shown in Figure 15, can be by between two terrestrial references L1 and L2 in the case where considering the track 105 of vehicle Terrestrial reference viewing angle φ is expressed as:
φ (x, a, L)=180 °-arctan (x/a)+arctan ((L-x)/a))
Therefore, worst error DmaxWith x, a, L is relevant.When shown in Figure 15 or traveling can be by following The principle of optimality draws terrestrial reference L1 and L2 optimization constellation, i.e. ratio a/L:
Figure 16 shows the optimization chart on terrestrial reference constellation according to one embodiment.In other words, Figure 16 shows a use In the example of the constellation for being applied to positioning of optimization road sign.Here, it is described optimization especially in conjunction with come from Fig. 1 determining device with And come from Fig. 3 for the determination method and/or come from Fig. 2 positioner and come from Fig. 4 be used for the positioning Method.The optimization can also be carried out in the case where using the principle of optimality described in reference diagram 15.
Here, optimization chart shows to come from Figure 15 kinematic parameter x on axis of abscissas, wherein, will be with kinematic parameter x phases The terrestrial reference viewing angle φ in units of degree closed is represented on axis of ordinates.One group 1600 of curve is shown in optimization chart Scheme or multiple curve maps for different ratio a/L are shown.
Figure 17 shows the optimization chart on terrestrial reference constellation according to one embodiment.Here, the chart in Figure 17 is corresponded to Come from Figure 16 chart, in addition to following situations, by related to kinematic parameter x by worst error DmaxMissed with terrestrial reference constellation The ratio of poor E formation is represented on axis of ordinates.Another group 1700 of the curve for different ratio a/L is shown in Figure 17 Figure.
Figure 18 shows the optimization chart on terrestrial reference constellation according to one embodiment.In other words, Figure 18 shows a use In the example of the constellation for being applied to positioning of optimization road sign.Here, it is described optimization especially in conjunction with come from Fig. 1 determining device with And come from Fig. 3 for the determination method and/or come from Fig. 2 positioner and come from Fig. 4 be used for the positioning Method.The optimization can also be carried out in the case where using the principle of optimality described in reference diagram 15.
Here, the optimization chart shows ratio a/L on axis of abscissas, wherein, by expression formula Represent on axis of ordinates.Curve map 1810 is shown in optimization chart, the curve map as shown in arrow in a/L etc. There is minimum value when about 0.4.Thus the particularly advantageous terrestrial reference constellation for a/L=0.4 is used.
Summarize and stated with other sentence again below with reference to Fig. 1 to 18 pairs of embodiment and background.
Multiple systems for highly automated traveling, robot and augmented reality are for example surveyed increasingly based on to high-precision The terrestrial reference LM, L1, L2 of amount analysis.It is possible thereby to determine the accurate location of vehicle 200 or robot in boundary coordinate system WKS. Particularly terrestrial reference LM, L1, L2 to the constellation and distance of vehicle 200 be critically important for that how to carry out positioning exactly.Mesh Preceding system, such as Video-SLAM methods (SLAM=Simultaneous Localization and Mapping, immediately Positioning and map structuring) need to analyze multiple terrestrial references, the terrestrial reference then can for example pass through beam compensation calculation out position.
According to one embodiment there is provided a kind of system or method for being used to optimally determine terrestrial reference LM, L1, L2, wherein, root Optimize the measure of terrestrial reference LM, L1, L2 and its constellation according to the track 105 or geometrical relationship of the Digital Highway map at place.Not only Terrestrial reference LM, the L1 are being surveyed and drawn, during L2 and in actual automatic running, the system can be used and help to reduce institute The terrestrial reference LM, L1, L2 to be stored that need quantity and the quantity for the terrestrial reference LM, L1, L2 to be analyzed.Also provide a kind of for above-mentioned The software product of system or the method for map provider, are also used for other application, the method for such as augmented reality system, The augmented reality system needs to position exactly to mix the virtual visual field and the true visual field.
In order to improve the precision of the positioning based on terrestrial reference, can reach for the positioning based on terrestrial reference can be calculated according to embodiment Precision.Highly automated traveling particularly requires to be positioned exactly with the order of magnitude of such as decimetre.Typical positioning side Method supports high-precision 3D terrestrial references.It can determine how accurate each terrestrial reference LM, L1, L2 position should be according to embodiment , so as to the precision needed for obtaining, it is necessary to how many terrestrial references LM, L1, L2, and/or how to realize favourable terrestrial reference geometry distribution.
For calculate the positioning based on terrestrial reference accessible precision process for example including to suitable sensor coordinates The selection of system, over the ground mark position error modelling, in 2D top views simplify geometric error model, according to terrestrial reference LM, L1, L2 precision, the viewing angle between terrestrial reference LM, L1, L2 and the measurement accuracy in the plane of delineation are carried out to site error General description, the inspection to the positioning scenarios in the highway scene for highly automated traveling and suitable for generating The step of terrestrial reference is distributed, as shown in Fig. 1 to 18.
If embodiment includes the "and/or" connection between fisrt feature and second feature, this can be read as, and implement Example not only has fisrt feature according to an embodiment but also with second feature and according to an other embodiment Or only there is fisrt feature or only there is second feature.

Claims (10)

1. a kind of terrestrial reference (L1, L2 for being used to determine to be used to position the object (200) of motion;LM method (300)), its In, this method has following step:
Use measure rule (140;542) generate (310) at least one constellation data with landmark data (130) and record (150), its Described in landmark data (130) represent terrestrial reference (L1, the L2 in the surrounding environment of the object (200) that are obtained by sensor; LM), wherein, it is described determine rule (140;542) it can be used, be applied to be determined from the landmark data (130) At least one pair of terrestrial reference (L1, the L2 of positioning;LM), wherein, track (105) and described pair of the first ground in the object (200) Mark the horizontal spacing (a) between (L1) and the terrestrial reference spacing between described pair of the first terrestrial reference (L1) and the second terrestrial reference (L2) (L) ratio between is in predetermined fit value scope, wherein, at least one constellation data record (150) has suitable For positioning a pair of terrestrial references (L1, L2 relative to the object (200);LM the position data of geometry constellation).
2. according to the method described in claim 1 (300), with the extraction landmark data (130) from original landmark data (532) The step of (320), the landmark data have it is unrelated with visibility conditions find value again, the value that finds again is more than most Small to find value again, the original terrestrial reference tables of data shows the total amount of terrestrial reference in the surrounding environment of the object (200) (L1, L2;LM), wherein, use the step of extracted landmark data (130) implements the generation (310).
3. according to any method of the preceding claims (300), with the use measure rule (140;542) And/or consider the landmark data (130), the position (536) of the object (200), the track (105 of the object (200); 534) and/or the object (200) speed (x) make a reservation for the fit value scope the step of (330).
4. according to any method of the preceding claims (300), with the constellation that will be generated from least one At least one pair of terrestrial reference (L1, the L2 that are applied to positioning of data record (150);LM position data) is added to numerical map (555) the step of map datum (550) (340).
5. a kind of method (400) for being used to position the object (200) of motion, wherein methods described (400) have following Step:
(410) at least one constellation data record (150) is read, the constellation data record is represented according in preceding claims The data record of method (300) generation described in any one;
Consider the terrestrial reference (L1, the L2 that are applied to positioning from least one constellation data record (150);LM positional number) According to and identification (420) terrestrial reference (L1, L2 in the image of the surrounding environment of the object (200);LM);And
Use terrestrial reference (L1, the L2 recognized in (420) the step of the identification;LM position data) is fixed to implement (430) Position.
6. method (400) according to claim 5, wherein the step of the implementation in (430), using the terrestrial reference number According to (130), the position (536) of the object (200), the object (200) track (105;And/or the object 534) (200) speed (x) implements the positioning.
7. the method (400) according to any one of claim 5 to 6, with the surrounding environment for obtaining the object (200) Image the step of (440).
8. one kind is arranged for carrying out according to any method of the preceding claims (300;400) device (110; 210)。
9. a kind of computer program, it is arranged for performing according to any method of the preceding claims (300; 400)。
10. a kind of machine readable storage medium, be stored with computer according to claim 9 on said storage Program.
CN201610941150.1A 2015-10-26 2016-10-25 Method and device for determining landmarks and method and device for positioning Active CN107044853B (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102015220831.1A DE102015220831A1 (en) 2015-10-26 2015-10-26 A method and apparatus for determining landmarks for a location determination for a moving object and method and apparatus for locating a moving object
DE102015220831.1 2015-10-26

Publications (2)

Publication Number Publication Date
CN107044853A true CN107044853A (en) 2017-08-15
CN107044853B CN107044853B (en) 2021-06-15

Family

ID=58490179

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610941150.1A Active CN107044853B (en) 2015-10-26 2016-10-25 Method and device for determining landmarks and method and device for positioning

Country Status (2)

Country Link
CN (1) CN107044853B (en)
DE (1) DE102015220831A1 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109862509A (en) * 2019-03-27 2019-06-07 合肥科塑信息科技有限公司 A kind of support WLAN fingerprint location sensing node positioning system
CN111089568A (en) * 2019-12-25 2020-05-01 上海点甜农业专业合作社 Road sign calibration instrument based on RTK + camera

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102017222810A1 (en) 2017-12-14 2019-06-19 Robert Bosch Gmbh Method for creating a feature-based localization map for a vehicle taking into account characteristic structures of objects
DE102018209366B4 (en) 2018-06-12 2020-07-02 Audi Ag Method for determining a position and / or orientation of a device
DE102019206918B3 (en) * 2019-05-13 2020-10-08 Continental Automotive Gmbh Position determination method and position determination device
DE102019003473A1 (en) 2019-05-16 2020-01-02 Daimler Ag Process for reliable vehicle localization using constellation code tables

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH09218712A (en) * 1996-02-08 1997-08-19 Murata Mach Ltd Guide system of unmanned vehicle
JP2007267231A (en) * 2006-03-29 2007-10-11 Toyota Motor Corp Optical axis offset detecting device of on-board camera and optical axis offset detecting method of the on-board camera
CN101762274A (en) * 2010-02-01 2010-06-30 北京理工大学 Observation condition number-based method for selecting autonomously located road sign of deep space probe
US20110010023A1 (en) * 2005-12-03 2011-01-13 Kunzig Robert S Method and apparatus for managing and controlling manned and automated utility vehicles
CN102442311A (en) * 2010-10-06 2012-05-09 罗伯特·博世有限公司 Method and device for determining processed image data about a sourround field of a vehicle
CN103562745A (en) * 2011-04-21 2014-02-05 科恩起重机有限公司 Techniques for positioning a vehicle
CN104457742A (en) * 2014-12-05 2015-03-25 歌尔声学股份有限公司 Target positioning method and target positioning equipment of object

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH09218712A (en) * 1996-02-08 1997-08-19 Murata Mach Ltd Guide system of unmanned vehicle
US20110010023A1 (en) * 2005-12-03 2011-01-13 Kunzig Robert S Method and apparatus for managing and controlling manned and automated utility vehicles
JP2007267231A (en) * 2006-03-29 2007-10-11 Toyota Motor Corp Optical axis offset detecting device of on-board camera and optical axis offset detecting method of the on-board camera
CN101762274A (en) * 2010-02-01 2010-06-30 北京理工大学 Observation condition number-based method for selecting autonomously located road sign of deep space probe
CN102442311A (en) * 2010-10-06 2012-05-09 罗伯特·博世有限公司 Method and device for determining processed image data about a sourround field of a vehicle
CN103562745A (en) * 2011-04-21 2014-02-05 科恩起重机有限公司 Techniques for positioning a vehicle
CN104457742A (en) * 2014-12-05 2015-03-25 歌尔声学股份有限公司 Target positioning method and target positioning equipment of object

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
DAVID SINRIECH 等: "Landmark con®guration for absolute positioning of autonomous vehicles", 《IIE TRANSACTIONS》 *
刘振宇 等: "基于人工路标和立体视觉的移动机器人自定位", 《计算机工程与应用》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109862509A (en) * 2019-03-27 2019-06-07 合肥科塑信息科技有限公司 A kind of support WLAN fingerprint location sensing node positioning system
CN111089568A (en) * 2019-12-25 2020-05-01 上海点甜农业专业合作社 Road sign calibration instrument based on RTK + camera

Also Published As

Publication number Publication date
CN107044853B (en) 2021-06-15
DE102015220831A1 (en) 2017-04-27

Similar Documents

Publication Publication Date Title
CN107044853A (en) Method and apparatus for definitely calibration method and device and for positioning
Ravi et al. Simultaneous system calibration of a multi-lidar multicamera mobile mapping platform
US20160209846A1 (en) Visual Localization Within LIDAR Maps
Wang et al. Automated road sign inventory system based on stereo vision and tracking
US20090154793A1 (en) Digital photogrammetric method and apparatus using intergrated modeling of different types of sensors
CN111436216A (en) Method and system for color point cloud generation
US20190072392A1 (en) System and method for self-geoposition unmanned aerial vehicle
US10930013B2 (en) Method and system for calibrating imaging system
JP2009294214A (en) Method and system for navigation based on topographic structure
CN101563581A (en) Method and apparatus for identification and position determination of planar objects in images
CN109146958B (en) Traffic sign space position measuring method based on two-dimensional image
CN111383205B (en) Image fusion positioning method based on feature points and three-dimensional model
CN108810473A (en) A kind of method and system for realizing GPS mapping camera views coordinates on a mobile platform
US10997785B2 (en) System and method for collecting geospatial object data with mediated reality
Anai et al. Automatic exterior orientation procedure for low-cost UAV photogrammetry using video image tracking technique and GPS information
CN114820793A (en) Target detection and target point positioning method and system based on unmanned aerial vehicle
JP2011112556A (en) Search target position locating device, method, and computer program
Mulsow et al. Subaquatic digital elevation models from UAV-imagery
US11513211B2 (en) Environment model using cross-sensor feature point referencing
WO2014112909A1 (en) Method and system for geo-referencing at least one sensor image
Hrabar et al. PTZ camera pose estimation by tracking a 3D target
Trusheim et al. Cooperative image orientation considering dynamic objects
KR100586815B1 (en) System and method for measuring position of threedimensions
JP7467722B2 (en) Feature Management System
KR102616437B1 (en) Method for calibration of lidar and IMU, and computer program recorded on record-medium for executing method therefor

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
GR01 Patent grant
GR01 Patent grant