CN110132306A - The correcting method and system of vehicle location error - Google Patents

The correcting method and system of vehicle location error Download PDF

Info

Publication number
CN110132306A
CN110132306A CN201910421222.3A CN201910421222A CN110132306A CN 110132306 A CN110132306 A CN 110132306A CN 201910421222 A CN201910421222 A CN 201910421222A CN 110132306 A CN110132306 A CN 110132306A
Authority
CN
China
Prior art keywords
pose
winding
vehicle
current
vision
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
CN201910421222.3A
Other languages
Chinese (zh)
Other versions
CN110132306B (en
Inventor
周建
潘力澜
李良
柴文楠
刘中元
林光模
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.)
Guangzhou Xiaopeng Motors Technology Co Ltd
Original Assignee
Guangzhou Xiaopeng Motors Technology Co Ltd
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 Guangzhou Xiaopeng Motors Technology Co Ltd filed Critical Guangzhou Xiaopeng Motors Technology Co Ltd
Priority to CN201910421222.3A priority Critical patent/CN110132306B/en
Publication of CN110132306A publication Critical patent/CN110132306A/en
Application granted granted Critical
Publication of CN110132306B publication Critical patent/CN110132306B/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
    • G01C21/005Navigation; 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
    • 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
    • G01C21/30Map- or contour-matching
    • G01C21/32Structuring or formatting of map data
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C25/00Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass

Abstract

The present invention relates to technical field of vehicle, the correcting method and system of a kind of vehicle location error are disclosed, it include: when detecting that winding phenomenon occurs in the process of moving in current vehicle, it will be merged with the matched winding vision pose of winding phenomenon and with the matched vehicle body pose of winding phenomenon, to generate global fusion map, wherein, winding vision pose is acquired by the visual sensor of current vehicle, and vehicle body pose is acquired by the IMU of current vehicle;Based on all key frames for including in the global fixed winding vision pose of fusion map, and the visual signature point determined from key frame is optimized, the Target key frames after being optimized;Winding detection is carried out to Target key frames, obtains the first winding testing result;First winding testing result is fed back into IMU, so that IMU carries out error correction to vehicle body pose.Implement the embodiment of the present invention, is able to ascend the accuracy of vehicle location function.

Description

The correcting method and system of vehicle location error
Technical field
The present invention relates to technical field of vehicle, and in particular to a kind of correcting method and system of vehicle location error.
Background technique
With the fast development of autonomous driving vehicle, positioning function almost becomes the necessary functions of autonomous driving vehicle. Currently, some autonomous driving vehicles can be according to the collected data building vehicle-periphery of sensor being arranged on vehicle SLAM map (Simultaneous Localization And Mapping, immediately positioning and building map), hereafter, automatically Driving can be by the Inertial Measurement Unit (Inertial Measurement Unit, IMU) of vehicle itself in preparatory structure It is positioned in the SLAM map built.
However, it has been found in practice that the data that vehicle is acquired usually using IMU were positioned in SLAM map, by There may be errors for the measuring instrument for including in IMU, therefore the collected data of IMU are also typically present error, so as to cause The accuracy of the positioning function of autonomous driving vehicle is poor.
Summary of the invention
The embodiment of the present invention discloses the correcting method and system of a kind of vehicle location error, is able to ascend vehicle location function Accuracy.
First aspect of the embodiment of the present invention discloses a kind of correcting method of vehicle location error, which comprises
It, will be with the matched winding of winding phenomenon when detecting that winding phenomenon occurs in the process of moving in current vehicle It vision pose and is merged with the matched vehicle body pose of the winding phenomenon, to generate global fusion map, wherein described Winding vision pose is acquired by the visual sensor of the current vehicle, and the vehicle body pose passes through the current vehicle IMU acquisition;
Based on the key frame in the global fixed winding vision pose of fusion map, and to from the key frame Determining visual signature point optimizes, the Target key frames after being optimized;
Winding detection is carried out to the Target key frames, obtains the first winding testing result;
The first winding testing result is fed back into the IMU, so that the IMU carries out error to the vehicle body pose It corrects.
As an alternative embodiment, in first aspect of the embodiment of the present invention, the method also includes:
When the current vehicle is in driving status, the Current vision pose of the visual sensor acquisition is obtained;
Global optimization is carried out to the Current vision pose, obtains Current vision optimization pose;
When detecting that Current vision optimization pose indicates that winding phenomenon occurs in the process of moving in the current vehicle When, winding detection is carried out to Current vision optimization pose, obtains the second winding testing result;
The second winding testing result is fed back into the IMU, to correct the running error of the IMU.
As an alternative embodiment, in first aspect of the embodiment of the present invention, it is described to the Current vision position Appearance carries out global optimization, obtains Current vision optimization pose, comprising:
Pass through the current key frame in the fixed Current vision pose of global optimization;
Optimize the visual signature point in the current key frame, pose is optimized with the Current vision after being optimized.
As an alternative embodiment, in first aspect of the embodiment of the present invention, it is described to the Target key frames Winding detection is carried out, the first winding testing result is obtained, comprising:
It obtains and several matched history keyword frames of the Target key frames;
Calculate the error of the Target key frames Yu each history keyword frame;
Overall error is obtained according to each error calculation, and the overall error is determined as the first winding testing result.
As an alternative embodiment, in first aspect of the embodiment of the present invention, it is described to detect current vehicle When occurring winding phenomenon in the process of moving, it will show with the matched winding vision pose of the winding phenomenon and with the winding As matched vehicle body pose is merged, before generating global fusion map, the method also includes:
The environmental visual pose of the current vehicle in the process of moving is acquired by the visual sensor;
When the current environment vision pose and any one of environmental visual position for detecting the visual sensor acquisition When appearance is similar, determine that winding phenomenon occurs in the process of moving in the current vehicle.
Second aspect of the embodiment of the present invention discloses the correcting method of another vehicle location error, which comprises
When vehicle is in driving status, the target visual pose of the visual sensor acquisition of the vehicle is obtained;
Pass through the key frame in the fixed target visual pose of global optimization;
Optimize the visual signature point in the key frame, pose is optimized with the target visual after being optimized;
It is right when detecting that the target visual optimization pose indicates that winding phenomenon occurs in the process of moving in the vehicle The target visual optimization pose carries out winding detection, obtains third winding result;
The third winding result is fed back to the IMU of the vehicle, to correct the running error of the IMU.
The third aspect of the embodiment of the present invention discloses a kind of correcting system of vehicle location error, comprising:
Associative cell will be with the winding for when detecting that winding phenomenon occurs in the process of moving in current vehicle It the matched winding vision pose of phenomenon and is merged with the matched vehicle body pose of the winding phenomenon, to generate global fusion Map, wherein the winding vision pose is acquired by the visual sensor of the current vehicle, and the vehicle body pose passes through institute State the IMU acquisition of current vehicle;
First optimization unit, for fixing the key frame in the winding vision pose based on the global fusion map, And the visual signature point determined from the key frame is optimized, the Target key frames after being optimized;
First detection unit obtains the first winding testing result for carrying out winding detection to the Target key frames;
First feedback unit, for the first winding testing result to be fed back to the IMU, so that the IMU is to institute It states vehicle body pose and carries out error correction.
As an alternative embodiment, in the third aspect of the embodiment of the present invention, the system also includes:
First acquisition unit, for when the current vehicle is in driving status, obtaining the visual sensor acquisition Current vision pose;
Second optimization unit obtains Current vision optimization pose for carrying out global optimization to the Current vision pose;
Second detection unit detects that the Current vision optimization pose indicates that the current vehicle is being run over for working as When occurring winding phenomenon in journey, winding detection is carried out to Current vision optimization pose, obtains the second winding testing result;
Second feedback unit, for the second winding testing result to be fed back to the IMU, to correct the IMU fortune Error in row.
As an alternative embodiment, in the third aspect of the embodiment of the present invention, the second optimization unit includes:
Fixed subelement, for passing through the current key frame in the fixed Current vision pose of global optimization;
Optimize subelement, for optimizing the visual signature point in the current key frame, to work as forward sight after being optimized Feel optimization pose.
As an alternative embodiment, in the third aspect of the embodiment of the present invention, the first detection unit includes:
Subelement is obtained, for obtaining and several matched history keyword frames of the Target key frames;
Computation subunit, for calculating the error of the Target key frames Yu each history keyword frame;
It determines subelement, for obtaining overall error according to each error calculation, and the overall error is determined as One winding testing result.
As an alternative embodiment, in the third aspect of the embodiment of the present invention, the system also includes:
Acquisition unit, for the associative cell will with the matched winding vision pose of the winding phenomenon and with institute It states the matched vehicle body pose of winding phenomenon to be merged, before generating global fusion map, be adopted by the visual sensor Collect the environmental visual pose of the current vehicle in the process of moving;
Determination unit, for when the current environment vision pose and any one institute for detecting the visual sensor acquisition State environmental visual pose it is similar when, determine that winding phenomenon occurs in the process of moving in the current vehicle.
Fourth aspect of the embodiment of the present invention discloses the correcting system of another vehicle location error, comprising:
Second acquisition unit, what the visual sensor for when vehicle is in driving status, obtaining the vehicle acquired Target visual pose;
Fixed cell, for passing through the key frame in the fixed target visual pose of global optimization;
Third optimizes unit, for optimizing the visual signature point in the key frame, with the target visual after being optimized Optimize pose;
Third detection unit detects that the target visual optimization pose indicates the vehicle in the process of moving for working as When there is winding phenomenon, winding detection is carried out to target visual optimization pose, obtains third winding result;
Third feedback unit, for the third winding result to be fed back to the IMU of the vehicle, to correct the IMU Running error.
The 5th aspect of the embodiment of the present invention discloses a kind of vehicle electronic device, comprising:
It is stored with the memory of executable program code;
The processor coupled with the memory;
The processor calls the executable program code stored in the memory, executes the embodiment of the present invention the The instruction of some or all of correcting method of vehicle location error step disclosed in one side;
Alternatively, the processor calls the executable program code stored in the memory, it is real to execute the present invention Apply the instruction of some or all of correcting method of vehicle location error step disclosed in a second aspect.
The 6th aspect of the embodiment of the present invention discloses a kind of vehicle, and the vehicle includes that the 5th aspect of the embodiment of the present invention is open Vehicle electronic device.
The 7th aspect of the embodiment of the present invention discloses a kind of computer readable storage medium, the computer readable storage medium Store program code, wherein said program code includes any one method for executing first aspect and second aspect Some or all of step instruction.
Eighth aspect of the embodiment of the present invention discloses a kind of computer program product, when the computer program product is calculating When being run on machine, so that the computer executes some or all of any one method of first aspect and second aspect step Suddenly.
The aspect of the embodiment of the present invention the 9th disclose a kind of using distribution platform, and the application distribution platform is for publication calculating Machine program product, wherein when the computer program product is run on computers, so that the computer executes first party Some or all of any one of face and second aspect method step.
Compared with prior art, the embodiment of the present invention has the advantages that
In the embodiment of the present invention, when detecting that winding phenomenon occurs in the process of moving in current vehicle, it will show with winding It is associated as matched winding vision pose and with the matched vehicle body pose of winding phenomenon, wherein winding vision pose is logical The visual sensor acquisition of current vehicle is crossed, vehicle body pose is acquired by the IMU of current vehicle;In fixed winding vision pose Key frame, and the visual signature point determined from key frame is optimized, the Target key frames after being optimized;Target is closed Key frame carries out winding detection, obtains the first winding testing result;First winding testing result is fed back into IMU, so that IMU is to vehicle Body pose carries out error correction.As it can be seen that implementing the embodiment of the present invention, the winding phenomenon in vehicle travel process can be being detected When, the corresponding winding vision pose of winding phenomenon is associated with vehicle body pose, and then is calculated according to winding vision pose The winding testing result arrived carries out error correction to collected vehicle body pose, to eliminate error existing for vehicle body pose, thus Improve the accuracy of vehicle location function.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to needed in the embodiment Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for ability For the those of ordinary skill of domain, without creative efforts, it can also be obtained according to these attached drawings other attached Figure.
Fig. 1 is a kind of flow diagram of the correcting method of vehicle location error disclosed by the embodiments of the present invention;
Fig. 2 is the flow diagram of the correcting method of another vehicle location error disclosed by the embodiments of the present invention;
Fig. 3 is the flow diagram of the correcting method of another vehicle location error disclosed by the embodiments of the present invention;
Fig. 4 is the flow diagram of the correcting method of another vehicle location error disclosed by the embodiments of the present invention;
Fig. 5 is a kind of structural schematic diagram of the correcting system of vehicle location error disclosed by the embodiments of the present invention;
Fig. 6 is the structural schematic diagram of the correcting system of another vehicle location error disclosed by the embodiments of the present invention;
Fig. 7 is the structural schematic diagram of the correcting system of another vehicle location error disclosed by the embodiments of the present invention;
Fig. 8 is the structural schematic diagram of the correcting system of another vehicle location error disclosed by the embodiments of the present invention;
Fig. 9 is a kind of structural schematic diagram of vehicle electronic device disclosed by the embodiments of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete Site preparation description, it is clear that the described embodiment is only a part of the embodiment of the present invention, instead of all the embodiments.Based on this Embodiment in invention, every other reality obtained by those of ordinary skill in the art without making creative efforts Example is applied, shall fall within the protection scope of the present invention.
It should be noted that term " includes " and " having " and their any changes in the embodiment of the present invention and attached drawing Shape, it is intended that cover and non-exclusive include.Such as contain the process, method of a series of steps or units, system, product or Equipment is not limited to listed step or unit, but optionally further comprising the step of not listing or unit or optional Ground further includes the other step or units intrinsic for these process, methods, product or equipment.
The embodiment of the present invention discloses the correcting method and system of a kind of vehicle location error, can be according to winding testing result Error correction is carried out to collected vehicle body pose, to eliminate error existing for vehicle body pose, to improve vehicle location function The accuracy of energy.It is described in detail separately below.
Embodiment one
Referring to Fig. 1, Fig. 1 is a kind of process signal of the correcting method of vehicle location error disclosed by the embodiments of the present invention Figure.As shown in Figure 1, the correcting method of the vehicle location error may comprise steps of:
101, when detecting that winding phenomenon occurs in the process of moving in current vehicle, vehicle electronic device will show with winding It is merged as matched winding vision pose and with the matched vehicle body pose of winding phenomenon, to generate global fusion map, Wherein, which is acquired by the visual sensor of current vehicle, which is adopted by the IMU of current vehicle Collection.
In the embodiment of the present invention, winding phenomenon can turn again to previous traveling warp for vehicle after travelling a period of time The place crossed.The visual sensor and IMU of current vehicle can be to be run based on the SLAM constructed in advance.Current vehicle Visual sensor may include image capture device (such as video camera) and visual odometry (Visual Odometry, VO), Image capture device can acquire the image of current vehicle in the process of moving, and visual odometry can be estimated based on SLAM map Movement relation between adjacent image, and then the motion state of image capture device is obtained, so that visual odometry can root The vision pose of current vehicle is determined according to the motion state of image capture device, and is showed when winding occurs in the process of moving in vehicle As when, which can be determined as to winding vision pose.Furthermore, it is possible to by winding in the SLAM map constructed in advance It vision pose and is associated with the matched vehicle body pose of winding phenomenon, to generate based on the SLAM map constructed in advance Global fusion map.
In the embodiment of the present invention, current vehicle can be got by IMU the travel speed of current vehicle, driving direction, The information such as acceleration and angular speed, vehicle electronic device can be according to the running condition information meters of the IMU current vehicle obtained Calculation obtains the vehicle body pose of current vehicle.
102, vehicle electronic device is based on the key frame in the global fixed winding vision pose of fusion map, and to from key The visual signature point determined in frame optimizes, the Target key frames after being optimized.
In the embodiment of the present invention, vehicle electronic device can be by the mode of pose figure (Pose Graph, PG) from winding It fixes all key frames in vision pose, and then global bundle adjustment (Global Bundle can be used Adjustment, GBA) all key frames are optimized, to obtain Target key frames.Wherein, vehicle electronic device acquires To winding vision pose in may include the collected several frame images of image capture device, therefore vehicle electronic device can be with Target image is determined from several frame images by way of PG, and the target image can be determined as key frame.
103, vehicle electronic device carries out winding detection to Target key frames, obtains the first winding testing result.
In the embodiment of the present invention, when detecting that winding phenomenon occurs in vehicle, it is believed that vehicle is not to travel for the first time To current location, it can think vehicle repeatedly by current location, and then can consider vehicle collect multiple current locations Corresponding vision pose.In addition, vehicle electronic device can carry out Target key frames by vision winding and scale re-optimization Winding detection, and then global optimization is carried out to overall situation fusion map according to the detection of this winding, so that obtained winding detection As a result more accurate.
As an alternative embodiment, vehicle electronic device carries out winding detection to Target key frames, first is obtained The mode of winding testing result may include following steps:
The initial visual position acquired for the first time is chosen in the corresponding vision pose in the current location of vehicle electronic device from acquisition Appearance, and initial key frame can be determined from the initial visual pose;
Vehicle electronic device can determine target visual characteristic point from Target key frames, can also be from initial key frame In determine initial visual characteristic point, wherein visual signature point can by key point (Key-point) and description son (Descriptor) two parts form;
Available target critical point and goal description in target visual characteristic point of vehicle electronic device, and can To get the initial key point in initial visual characteristic point and initial description;
Vehicle electronic device can compare the first difference between target critical point and initial key point, can also compare mesh The second difference between mark description and initial description;
The first winding testing result comprising the first difference and the second difference can be generated in vehicle electronic device.
Wherein, implement this embodiment, it can be by current collected Target key frames and the initial view acquired in the past Feel that the initial visual characteristic point of pose compares, so that the first winding testing result of this winding detection is obtained, to mention The high accuracy of the first winding testing result of output.
It may include several pixels in the embodiment of the present invention, in key point, vehicle electronic device can determine each picture The color of vegetarian refreshments can also determine that the layout of each pixel in key point determines between each pixel in key point Positional relationship, and then the colour-difference of pixel between target critical point and initial key point can be compared by vehicle electronic device Placement differences between different and each pixel, to obtain the first difference comprising color difference and placement differences;This Outside, vehicle electronic device can obtain several region units around characteristic point, and calculate each region block and characteristic point Harr wavelet character obtains the principal direction of characteristic point region, and then vehicle electronics from according to several Harr wavelet characters It is exhausted relative to the horizontal direction value, vertical direction value, horizontal direction of principal direction that equipment can calculate the corresponding key point of characteristic point To value and vertical direction absolute value, and can determine comprising horizontal direction value, vertical direction value, horizontal direction absolute value with And vertical direction absolute value description son, it is seen then that vehicle electronic device can by the horizontal direction value of goal description with initially The horizontal direction value of description compares to obtain that horizontal direction value difference is different, by the vertical direction value of goal description with initially retouch The vertical direction value for stating son compares to obtain that vertical direction value difference is different, by the horizontal direction absolute value of goal description with initially The horizontal direction absolute value of description compares to obtain horizontal direction absolute value difference, and the vertical direction of goal description is exhausted The vertical direction absolute value of value and initial description is compared to obtain vertical direction absolute value difference, to obtain comprising water It is square different to value difference, vertical direction value difference is different, horizontal direction absolute value difference and vertical direction absolute value difference it is second poor It is different.
104, the first winding testing result is fed back to IMU by vehicle electronic device, so that IMU carries out error to vehicle body pose It corrects.
As an alternative embodiment, the first winding testing result is fed back to IMU by vehicle electronic device, so that The mode that IMU carries out error correction to vehicle body pose may include following steps:
Vehicle electronic device obtains location information corresponding with the first winding testing result;
Vehicle electronic device obtains the matched basic body pose of the location information;
Vehicle electronic device analyzes the vehicle body pose of IMU acquisition and the vehicle body position and attitude error of the basic body pose;
The vehicle body position and attitude error is fed back to IMU by vehicle electronic device, so that IMU carries out error correction to vehicle body pose.
Wherein, implement this embodiment, vehicle electronic device can be made to be passed according to the first winding testing result to vision Sensor carries out error correction, to obtain the standard vision pose of current vehicle present position, and then can obtain from SLAM map Location information of the current vehicle corresponding with standard vision pose in the SLAM map is got, so that vehicle electronic device can be with Get with the matched basic body pose of the location information, since the IMU of vehicle electronic device may be deposited in position fixing process The error accumulation the case where, therefore the collected vehicle body pose of the IMU of basic body pose and current vehicle can be carried out pair Than, the vehicle body position and attitude error of the vehicle body pose of IMU acquisition is obtained, and then the vehicle body position and attitude error can be fed back to IMU, so that IMU is the error accumulation occurred in position fixing process according to elimination IMU with vehicle body position and attitude error.
In the method depicted in fig. 1, error can be carried out to collected vehicle body pose according to winding testing result to entangle Just, to eliminate error existing for vehicle body pose, to improve the accuracy of vehicle location function.In addition, implementing described by Fig. 1 Method, improve output the first winding testing result accuracy.In addition, implement method described in Fig. 1, it can be to IMU The vehicle body position and attitude error is fed back, so that IMU is tired according to the error that IMU occurs in position fixing process is eliminated with vehicle body position and attitude error Product.
Embodiment two
Referring to Fig. 2, the process that Fig. 2 is the correcting method of another vehicle location error disclosed by the embodiments of the present invention is shown It is intended to.Compared with embodiment one, result that the IMU that the embodiment of the present invention increases vehicle is detected according to winding in IMU to existing The method corrected of error and the more detailed method for illustrating to carry out global optimization to Current vision pose, both The accuracy for correcting imu error is improved, and the effect of optimization to Current vision optimization pose can be improved.As shown in Fig. 2, should The correcting method of vehicle location error may comprise steps of:
Step 201~step 204 is identical as step 101~step 104, and the following contents repeats no more.
205, when current vehicle is in driving status, vehicle electronic device obtains the Current vision of visual sensor acquisition Pose.
In the embodiment of the present invention, the visual sensor of vehicle collected vision pose when being remained static due to vehicle It does not change, therefore, vehicle can Current vision pose of acquisition vehicle when being in driving status.
206, vehicle electronic device carries out global optimization to Current vision pose, obtains Current vision optimization pose.
In the embodiment of the present invention, global optimization can optimize for GBA, and vehicle electronic device can be by GBA to working as forward sight Feel that pose optimizes, it can optimize to the current key frame in Current vision pose, to obtain optimizing current pass The Current vision of key frame optimizes pose.
As an alternative embodiment, vehicle electronic device carries out global optimization to Current vision pose, worked as The mode of anterior optic optimization pose may include following steps:
Vehicle electronic device passes through the current key frame in the fixed Current vision pose of global optimization;
Vehicle electronic device optimizes the visual signature point in current key frame, optimizes position with the Current vision after being optimized Appearance.
Wherein, implement this embodiment, the key in Current vision pose can be fixed by way of global optimization Frame, and optimize the visual signature point in the key frame, so that the Current vision after being optimized optimizes pose, to improve to working as The effect of optimization of anterior optic optimization pose.
In the embodiment of the present invention, current key frame can be the key frame got from Current vision pose, due to vehicle The collected Current vision pose of visual sensor in may include several key frames, therefore vehicle electronic device can be from Any one frame current key frame is fixed in several key frames, and then GBA can be used, current key frame is optimized, to obtain The visual signature point of optimization in current key frame, vehicle electronic device can be by the current passes of the visual signature point comprising optimization The corresponding pose of key frame is determined as Current vision optimization pose.
207, when detecting that winding phenomenon occurs in the process of moving in Current vision optimization pose instruction current vehicle, vehicle It carries electronic equipment and winding detection is carried out to Current vision optimization pose, obtain the second winding testing result.
In the embodiment of the present invention, the vehicle electronic device of current vehicle first collected to visual sensor can work as forward sight Feel that pose optimizes, so that vehicle electronic device can be using the Current vision optimization pose after optimizing as foundation, to vehicle Winding phenomenon is detected, and the available previous collected Current vision of vehicle electronic device optimizes the corresponding position of pose History vision pose, and history keyword frame can be determined from the history vision pose;Vehicle electronic device can be from going through History visual signature point is determined in history key frame;Vehicle electronic device can also get the history in history visual signature point Key point and Reassessment, and the current key point in the available visual signature point to current key frame and currently retouch State son;Vehicle electronic device can compare the third difference between history keyword point and current key point, can also compare history The 4th difference between description and current description;Vehicle electronic device can be generated comprising third difference and the 4th difference Second winding testing result.
208, the second winding testing result is fed back to IMU by vehicle electronic device, to correct the running error of IMU.
In the embodiment of the present invention, implement above-mentioned step 205~step 208, it can be right when vehicle is in driving status The Current vision pose that visual sensor is got optimizes, and then optimizes pose by the Current vision after optimization and detect vehicle Currently whether there is winding phenomenon, and after detecting winding phenomenon to Current vision optimization pose carry out winding detection, and The result that winding detects is fed back into the IMU of vehicle, so that the result that the IMU of vehicle is detected according to winding is to present in IMU Error is corrected, to improve the accuracy of winding phenomenon detection, and then can promote the accuracy for correcting imu error.
In the embodiment of the present invention, before arbitrary steps of the step 205~step 208 in step 201~step 204 or It executes later and the embodiment of the present invention is not impacted.
In the method depicted in fig. 2, error can be carried out to collected vehicle body pose according to winding testing result to entangle Just, to eliminate error existing for vehicle body pose, to improve the accuracy of vehicle location function.In addition, implementing described by Fig. 2 Method, improve to Current vision optimization pose effect of optimization.In addition, implementing method described in Fig. 2, it is existing to improve winding As the accuracy of detection, and then the accuracy for correcting imu error can be promoted.
Embodiment three
Referring to Fig. 3, the process that Fig. 3 is the correcting method of another vehicle location error disclosed by the embodiments of the present invention is shown It is intended to.Compared with embodiment one, the embodiment of the present invention increases the method for winding phenomenon detection and more detailed illustrates The calculation method method of first winding result had not only improved the accuracy for the detection of vehicle winding phenomenon, but also can be improved back The accuracy of ring testing result.As shown in figure 3, the correcting method of the vehicle location error may comprise steps of:
301, vehicle electronic device acquires the environmental visual pose of current vehicle in the process of moving by visual sensor.
In the embodiment of the present invention, the visual sensor of vehicle may include image capture device and visual odometry, vehicle-mounted Electronic equipment can control several images of image capture device acquisition vehicle in the process of moving, and visual odometry can be with base The movement relation between SLAM map estimation arbitrary neighborhood image, and then analyzed to obtain image capture device according to movement relation Motion state, so that visual odometry can determine the ring of vehicle in the process of moving according to the motion state of image capture device Border vision pose.
302, when the current environment vision pose and any one environmental visual pose phase for detecting visual sensor acquisition Like when, vehicle electronic device determines that winding phenomenon occurs in the process of moving in current vehicle.
In the embodiment of the present invention, vehicle can repeatedly pass through same position in the process of moving, it can be considered that vehicle The corresponding vision pose in multiple current locations can be collected, vehicle electronic device can regard current collected current environment Feel that pose and previous collected environment pose carry out the calculating of similarity, to determine whether vehicle returns according to similarity Ring phenomenon.
In the embodiment of the present invention, current environment can be regarded by way of bag of words (Bag-of-words Model) Feel that each visual signature point for including in pose and the corresponding vision pose of current environment vision pose is determined as word (Word), And then generate the dictionary (Dictionary) comprising all words;Further, since different words (i.e. visual signature point) for The similarity judgement contribution of different key frames is different, therefore different words should have different weights;The weight can lead to Two amounts are crossed to describe: inverse document frequency (Inverse Document Frequency, IDF) and translating frequency (Term Frequency, TF), wherein IDF describes the frequency that word occurs when constructing dictionary, and the lower IDF the more representative;TF is retouched The frequency that word occurs in single-frame images is stated, the higher TF the more representative;The vision for including in any one frame vision pose The quantity of characteristic point can be N number of, and can be ranked up in advance to each visual signature point, so that each visual signature point is all Unique serial number i can be corresponded to.
Wherein, the calculation formula of IDF can be with are as follows:
N can indicate that the number for occurring each word in dictionary is added in above-mentioned formula, and obtained each word goes out Occurrence number summation, i can indicate the serial number (such as 1,2,3) for currently calculating the word of IDF, therefore, niIt can indicate serial number i The number that occurs in dictionary of word;
In addition, the calculation formula of TF can be with are as follows:
N, i and n in above-mentioned formulaiWith n, i and n in the calculation formula of IDFiThe meaning of expression is identical;Therefore it calculates The weight η of the word of serial number iiCalculation formula can be with are as follows:
ηi=TFi×IDFi
Therefore, a frame key frame A is by word niAnd the bag of words of word corresponding weight η composition can be with are as follows:
Wherein, υAFor single vector, due to the quantity of word be it is N number of, i can be the Any Digit in 1~N;Together When, it can also be by the ownership renormalization in a frame key frame:
When calculating the difference between two frame key frames, similarity can be obtained by comparing corresponding weight, when two frames When similarity between key frame is higher, it is believed that above-mentioned two frame is winding.
In the embodiment of the present invention, implement above-mentioned step 301~step 302, the visual sensor of vehicle can be acquired Current environment vision pose compared with previous collected environmental visual pose, to judge position that vehicle current driving is arrived The position whether before run over for it is set, if so, it is considered that winding phenomenon occurs in current driving process in vehicle, from And improve the accuracy detected for vehicle winding phenomenon.
303, when detecting that winding phenomenon occurs in the process of moving in current vehicle, vehicle electronic device will show with winding It is merged as matched winding vision pose and with the matched vehicle body pose of winding phenomenon, to generate global fusion map, Wherein, which is acquired by the visual sensor of current vehicle, which is adopted by the IMU of current vehicle Collection.
304, vehicle electronic device is based on the key frame in the global fixed winding vision pose of fusion map, and to from key The visual signature point determined in frame optimizes, the Target key frames after being optimized.
305, vehicle electronic device obtains and several matched history keyword frames of Target key frames.
306, vehicle electronic device calculates the error of Target key frames and each history keyword frame.
It, can be based on the pose figure in the primary pose figure of PG, g2o or Lie algebra to target critical in the embodiment of the present invention The error of frame and each history keyword frame is calculated, and the Target key frames based on SLAM map can be calculated and go through with each Error between history key frame.
307, vehicle electronic device obtains overall error according to each error calculation, and overall error is determined as the inspection of the first winding Survey result.
In the embodiment of the present invention, implement above-mentioned step 305~step 307, Target key frames can be calculated and gone through with each Error between history key frame, and overall error may further be calculated according to each error, obtained overall error can Be determined as winding detection as a result, so as to improve by several between Target key frames and several history keyword frames The accuracy for the winding testing result that a error obtains.
308, the first winding testing result is fed back to IMU by vehicle electronic device, so that IMU carries out error to vehicle body pose It corrects.
In the method depicted in fig. 3, error can be carried out to collected vehicle body pose according to winding testing result to entangle Just, to eliminate error existing for vehicle body pose, to improve the accuracy of vehicle location function.In addition, described by implementing Fig. 3 Method, improve the accuracy that detects for vehicle winding phenomenon.In addition, method described in implementing Fig. 3, can be improved and pass through The accuracy for the winding testing result that several errors between Target key frames and several history keyword frames obtain.
Example IV
Referring to Fig. 4, the process that Fig. 4 is the correcting method of another vehicle location error disclosed by the embodiments of the present invention is shown It is intended to.As shown in figure 4, the correcting method of the vehicle location error may comprise steps of:
401, when vehicle is in driving status, vehicle electronic device obtains the target view of the visual sensor acquisition of vehicle Feel pose.
In the embodiment of the present invention, the visual sensor of vehicle can be to be run based on the SLAM constructed in advance.Vehicle Visual sensor may include image capture device (such as video camera) and visual odometry, image capture device can be adopted Collecting the image of vehicle in the process of moving, visual odometry can estimate movement relation between adjacent image based on SLAM map, And then the motion state of image capture device is obtained, so that visual odometry can be according to the movement shape of image capture device State determines the target visual pose of vehicle.In addition, vehicle electronic device can be the electronic equipments such as vehicle-mounted computer, vehicle can be passed through It carries electronic equipment and server carries out data interaction, collected data can also be carried out by operation by vehicle electronic device, To realize the normal operation of vehicle electronic device.
402, vehicle electronic device passes through the key frame in the fixed target visual pose of global optimization.
In the embodiment of the present invention, global optimization can pass through PG's for the optimization of the mode of PG, vehicle electronic device Mode fixed any one frame key frame out from target visual pose, so that vehicle electronic device is subsequent excellent to key frame progress Change.
403, the visual signature point in vehicle electronic device optimization key frame optimizes position with the target visual after being optimized Appearance.
In the embodiment of the present invention, vehicle electronic device can be used GBA to key frame fixed in target visual pose into Row optimization, to obtain the target visual characteristic point of the optimization in the key frame, vehicle electronic device can be by the mesh comprising optimization The corresponding target visual pose of mark visual signature point is determined as target visual optimization pose.
404, when detecting that winding phenomenon occurs in the process of moving in target visual optimization pose instruction vehicle, vehicle mounted electric Sub- equipment carries out winding detection to target visual optimization pose, obtains third winding result.
In the embodiment of the present invention, target visual pose can be acquired in real time in vehicle travel process, and to the mesh of acquisition Mark vision pose is detected, if detecting that winding phenomenon occurs in vehicle, so that it may be returned according to the target visual pose Ring detection, and error correction timely is carried out to the IMU of vehicle in vehicle operation, so that vehicle in the process of running may be used To carry out winding detection at any time, and error correction is carried out to IMU at any time, improves the timeliness of error correction.
405, third winding result is fed back to the IMU of vehicle by vehicle electronic device, to correct the running error of IMU.
In the method described in Fig. 4, error can be carried out to collected vehicle body pose according to winding testing result and entangled Just, to eliminate error existing for vehicle body pose, to improve the accuracy of vehicle location function.
Embodiment five
Referring to Fig. 5, Fig. 5 is a kind of structural representation of the correcting system of vehicle location error disclosed by the embodiments of the present invention Figure.As shown in figure 5, the correcting system of the vehicle location error may include:
Associative cell 501, for that will show with winding when detecting that winding phenomenon occurs in the process of moving in current vehicle It is merged as matched winding vision pose and with the matched vehicle body pose of winding phenomenon, to generate global fusion map, Wherein, winding vision pose is acquired by the visual sensor of current vehicle, and vehicle body pose is acquired by the IMU of current vehicle.
First optimization unit 502, the global fixed winding vision pose of fusion map for being generated based on associative cell 501 In key frame, and to from key frame determine visual signature point optimize, the Target key frames after being optimized.
First detection unit 503, the Target key frames for obtaining to the first optimization unit 502 carry out winding detection, obtain To the first winding testing result.
As an alternative embodiment, first detection unit 503 carries out winding detection to Target key frames, the is obtained The mode of one winding testing result is specifically as follows:
Choose the initial visual pose acquired for the first time from the corresponding vision pose in the current location of acquisition, and can be from Initial key frame is determined in the initial visual pose;
Target visual characteristic point is determined from Target key frames, and initial visual can also be determined from initial key frame Characteristic point;
The target critical point and goal description in target visual characteristic point are got, and available to initial visual Initial key point and initial description in characteristic point;
The first difference between target critical point and initial key point is compared, goal description can also be compared and initially retouched State the second difference between son;
Generate the first winding testing result comprising the first difference and the second difference.
Wherein, implement this embodiment, it can be by current collected Target key frames and the initial view acquired in the past Feel that the initial visual characteristic point of pose compares, so that the first winding testing result of this winding detection is obtained, to mention The high accuracy of the first winding testing result of output.
First feedback unit 504, the first winding testing result for obtaining first detection unit 503 feed back to IMU, So that IMU carries out error correction to vehicle body pose.
As an alternative embodiment, the first winding testing result is fed back to IMU by the first feedback unit 504, with The mode for making IMU carry out error correction to vehicle body pose is specifically as follows:
Obtain location information corresponding with the first winding testing result;
Obtain the matched basic body pose of the location information;
Analyze the vehicle body pose of IMU acquisition and the vehicle body position and attitude error of the basic body pose;
The vehicle body position and attitude error is fed back into IMU, so that IMU carries out error correction to vehicle body pose.
Wherein, implement this embodiment, error can be carried out to visual sensor according to the first winding testing result and entangled Just, it to obtain the standard vision pose of current vehicle present position, and then can be got from SLAM map and standard vision Location information of the corresponding current vehicle of pose in the SLAM map, to get and the matched basic body of the location information The case where pose, there may be error accumulations in position fixing process due to the IMU of vehicle, can by basic body pose with The collected vehicle body pose of the IMU of current vehicle compares, and obtains the vehicle body position and attitude error of the vehicle body pose of IMU acquisition, into And the vehicle body position and attitude error can be fed back to IMU, so that IMU is that foundation eliminates IMU in position fixing process with vehicle body position and attitude error The error accumulation of appearance.
As it can be seen that implement the correcting system of vehicle location error described in Fig. 5, it can be according to winding testing result to acquisition The vehicle body pose arrived carries out error correction, to eliminate error existing for vehicle body pose, to improve the standard of vehicle location function True property.In addition, implementing the correcting system of vehicle location error described in Fig. 5, the first winding testing result of output is improved Accuracy.In addition, implementing the correcting system of vehicle location error described in Fig. 5, vehicle body pose mistake can be fed back to IMU Difference, so that IMU is the error accumulation occurred in position fixing process according to elimination IMU with vehicle body position and attitude error.
Embodiment six
Referring to Fig. 6, the structure that Fig. 6 is the correcting system of another vehicle location error disclosed by the embodiments of the present invention is shown It is intended to.Wherein, the correcting system of vehicle location error shown in fig. 6 is the correcting system of vehicle location error as shown in Figure 5 It optimizes.Compared with the correcting system of vehicle location error shown in fig. 5, vehicle location error shown in fig. 6 The side that the result that the IMU that correcting system further increases vehicle is detected according to winding corrects error present in IMU Method and the more detailed method for illustrating to carry out global optimization to Current vision pose, both improve and correct imu error Accuracy, and the effect of optimization to Current vision optimization pose, the correcting system of vehicle location error shown in fig. 6 can be improved Can also include:
First acquisition unit 505, for when current vehicle is in driving status, obtaining the current of visual sensor acquisition Vision pose.
Second optimization unit 506, the Current vision pose for obtaining to first acquisition unit 505 carry out global optimization, Obtain Current vision optimization pose.
Second detection unit 507, for when the Current vision optimization pose instruction for detecting that the second optimization unit 506 obtains When winding phenomenon occurs in the process of moving in current vehicle, winding detection is carried out to Current vision optimization pose, is obtained second time Ring testing result.
Second feedback unit 508, the second winding testing result for obtaining second detection unit 507 feed back to IMU, To correct the running error of IMU.
, can be when vehicle be in driving status in the embodiment of the present invention, Current vision that visual sensor is got Pose optimizes, and then optimizes pose detection vehicle by the Current vision after optimization and currently whether there is winding phenomenon, and Winding detection is carried out to Current vision optimization pose after detecting winding phenomenon, and the result that winding detects is fed back into vehicle IMU so that the IMU of vehicle corrects error present in IMU according to the result that winding detects, to improve winding The accuracy of phenomenon detection, and then the accuracy for correcting imu error can be promoted.
As an alternative embodiment, the second optimization unit of the correcting system of vehicle location error shown in fig. 6 506 may include:
Fixed subelement 5061, the Current vision pose for being obtained by the fixed first acquisition unit 505 of global optimization In current key frame;
Optimize subelement 5062, for optimizing the visual signature point in the fixed current key frame of fixed subelement 5061, Optimize pose with the Current vision after being optimized.
Wherein, implement this embodiment, the key in Current vision pose can be fixed by way of global optimization Frame, and optimize the visual signature point in the key frame, so that the Current vision after being optimized optimizes pose, to improve to working as The effect of optimization of anterior optic optimization pose.
As it can be seen that implement the correcting system of vehicle location error described in Fig. 6, it can be according to winding testing result to acquisition The vehicle body pose arrived carries out error correction, to eliminate error existing for vehicle body pose, to improve the standard of vehicle location function True property.In addition, implementing the correcting system of vehicle location error described in Fig. 6, the optimization to Current vision optimization pose is improved Effect.In addition, implementing the correcting system of vehicle location error described in Fig. 6, the accuracy of winding phenomenon detection is improved, into And the accuracy for correcting imu error can be promoted.
Embodiment seven
Referring to Fig. 7, the structure that Fig. 7 is the correcting system of another vehicle location error disclosed by the embodiments of the present invention is shown It is intended to.Wherein, the correcting system of vehicle location error shown in Fig. 7 is the correcting system of vehicle location error as shown in Figure 6 It optimizes.Compared with the correcting system of vehicle location error shown in fig. 6, vehicle location error shown in Fig. 7 Correcting system further increases the method for winding phenomenon detection and more detailed illustrates the calculating of the first winding result Method method had not only improved the accuracy for the detection of vehicle winding phenomenon, but also the accuracy of winding testing result can be improved, The first detection unit 503 of the correcting system of vehicle location error shown in Fig. 7 may include:
Obtain subelement 5031, for obtains with first optimize the obtained Target key frames of unit 502 it is matched several History keyword frame.
Computation subunit 5032, for calculating the Target key frames and obtain subelement that the first optimization unit 502 obtains The error of the 5031 each history keyword frames obtained.
Determining subelement 5033, each error calculation for being obtained according to computation subunit 5032 obtains overall error, and Overall error is determined as the first winding testing result.
In the embodiment of the present invention, the error between Target key frames and each history keyword frame can be calculated, and according to Overall error may further be calculated in each error, obtained overall error be assured that for winding detection as a result, to The winding testing result obtained by several errors between Target key frames and several history keyword frames can be improved Accuracy.
As an alternative embodiment, the correcting system of vehicle location error shown in Fig. 7 can also include:
Acquisition unit 509, being used for will be with the matched winding vision pose of winding phenomenon and and winding in associative cell 501 The matched vehicle body pose of phenomenon is merged, and before generating global fusion map, acquires current vehicle by visual sensor Environmental visual pose in the process of moving;
Determination unit 510, for being adopted when the current environment vision pose for detecting visual sensor acquisition with any one When the environmental visual pose that collection unit 509 acquires is similar, determine that winding phenomenon occurs in the process of moving in current vehicle.
Wherein, implement this embodiment, can to the visual sensor of vehicle acquire current environment vision pose with Previous collected environmental visual pose compared, with judge vehicle current driving to position whether before run over for it Position, if so, it is considered that winding phenomenon occurs in current driving process in vehicle, to improve existing for vehicle winding As the accuracy of detection.
As it can be seen that implement the correcting system of vehicle location error described in Fig. 7, it can be according to winding testing result to acquisition The vehicle body pose arrived carries out error correction, to eliminate error existing for vehicle body pose, to improve the standard of vehicle location function True property.In addition, implementing the correcting system of vehicle location error described in Fig. 7, the standard detected for vehicle winding phenomenon is improved True property.In addition, implement Fig. 7 described in vehicle location error correcting system, can be improved by Target key frames with it is several The accuracy for the winding testing result that several errors between a history keyword frame obtain.
Embodiment eight
Referring to Fig. 8, the structure that Fig. 8 is the correcting system of another vehicle location error disclosed by the embodiments of the present invention is shown It is intended to.As shown in figure 8, the correcting system of the vehicle location error may include:
Second acquisition unit 801, the mesh of the visual sensor acquisition for when vehicle is in driving status, obtaining vehicle Mark vision pose.
Fixed cell 802, in the target visual pose by the fixed acquisition of second acquisition unit 801 of global optimization Key frame.
Third optimizes unit 803, for optimizing the visual signature point in the fixed key frame of fixed cell 802, to obtain Target visual after optimization optimizes pose.
Third detection unit 804, for when the target visual optimization pose instruction for detecting that third optimization unit 803 obtains When winding phenomenon occurs in the process of moving in vehicle, winding detection is carried out to target visual optimization pose, obtains third winding knot Fruit.
Third feedback unit 805, the third winding result for obtaining third detection unit 804 feed back to vehicle IMU, to correct the running error of IMU.
As it can be seen that implement the correcting system of vehicle location error described in Fig. 8, it can be according to winding testing result to acquisition The vehicle body pose arrived carries out error correction, to eliminate error existing for vehicle body pose, to improve the standard of vehicle location function True property.
Embodiment nine
Referring to Fig. 9, Fig. 9 is a kind of structural schematic diagram of vehicle electronic device disclosed by the embodiments of the present invention.Such as Fig. 9 institute Show, which may include:
It is stored with the memory 901 of executable program code;
The processor 902 coupled with memory 901;
Wherein, processor 902 calls the executable program code stored in memory 901, executes above method embodiment Some or all of method in one~embodiment of the method three step;
Alternatively, processor 902 calls the executable program code stored in memory 901, above method embodiment is executed Some or all of method in four step.
A kind of vehicle is also disclosed in the embodiment of the present invention, wherein the vehicle includes the vehicle electronic device in embodiment nine.
A kind of computer readable storage medium is also disclosed in the embodiment of the present invention, wherein computer-readable recording medium storage Program code, wherein program code includes for executing some or all of the method in above each method embodiment step Instruction.
A kind of computer program product is also disclosed in the embodiment of the present invention, wherein when computer program product on computers When operation, so that computer executes some or all of the method in such as above each method embodiment step.
The embodiment of the present invention is also disclosed a kind of using distribution platform, wherein using distribution platform for issuing computer journey Sequence product, wherein when computer program product is run on computers, so that computer executes such as the above each method embodiment In some or all of method step.
It should be understood that " embodiment of the present invention " that specification is mentioned in the whole text mean special characteristic related with embodiment, Structure or characteristic is included at least one embodiment of the present invention.Therefore, the whole instruction occur everywhere " in the present invention In embodiment " not necessarily refer to identical embodiment.In addition, these a particular feature, structure, or characteristics can be with any suitable Mode combines in one or more embodiments.Those skilled in the art should also know that embodiment described in this description Alternative embodiment is belonged to, related actions and modules are not necessarily necessary for the present invention.
In various embodiments of the present invention, it should be appreciated that magnitude of the sequence numbers of the above procedures are not meant to execute suitable Successively, the execution sequence of each process should be determined by its function and internal logic the certainty of sequence, without coping with the embodiment of the present invention Implementation process constitutes any restriction.
In addition, the terms " system " and " network " are often used interchangeably herein.It should be understood that the terms "and/or", only a kind of incidence relation for describing affiliated partner, indicates may exist three kinds of relationships, such as A and/or B, can To indicate: individualism A exists simultaneously A and B, these three situations of individualism B.In addition, character "/" herein, typicallys represent Forward-backward correlation object is a kind of relationship of "or".
In embodiment provided by the present invention, it should be appreciated that " B corresponding with A " indicates that B is associated with A, can be with according to A Determine B.It is also to be understood that determine that B is not meant to determine B only according to A according to A, it can also be according to A and/or other information Determine B.
Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of above-described embodiment is can It is completed with instructing relevant hardware by program, which can be stored in a computer readable storage medium, storage Medium include read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), programmable read only memory (Programmable Read-only Memory, PROM), erasable programmable is read-only deposits Reservoir (Erasable Programmable Read Only Memory, EPROM), disposable programmable read-only memory (One- Time Programmable Read-Only Memory, OTPROM), the electronics formula of erasing can make carbon copies read-only memory (Electrically-Erasable Programmable Read-Only Memory, EEPROM), CD-ROM (Compact Disc Read-Only Memory, CD-ROM) or other disc memories, magnetic disk storage, magnetic tape storage or can For carrying or any other computer-readable medium of storing data.
Above-mentioned unit as illustrated by the separation member may or may not be physically separated, aobvious as unit The component shown may or may not be physical unit, can be in one place, or may be distributed over multiple nets On network unit.Some or all of units can be selected to achieve the purpose of the solution of this embodiment according to the actual needs.
In addition, each functional unit in various embodiments of the present invention can integrate in one processing unit, it is also possible to Each unit physically exists alone, and can also be integrated in one unit with two or more units.Above-mentioned integrated unit Both it can take the form of hardware realization, can also realize in the form of software functional units.
If above-mentioned integrated unit is realized in the form of SFU software functional unit and when sold or used as an independent product, It can store in a retrievable memory of computer.Based on this understanding, technical solution of the present invention substantially or Person says all or part of of the part that contributes to existing technology or the technical solution, can be in the form of software products It embodies, which is stored in a memory, including several requests are with so that a computer is set Standby (can be personal computer, server or network equipment etc., specifically can be the processor in computer equipment) executes Some or all of each embodiment above method of the invention step.
Detailed Jie has been carried out to the correcting method and system of a kind of vehicle location error disclosed by the embodiments of the present invention above It continues, used herein a specific example illustrates the principle and implementation of the invention, and the explanation of above embodiments is only It is to be used to help understand method and its core concept of the invention;At the same time, for those skilled in the art, according to this hair Bright thought, there will be changes in the specific implementation manner and application range, in conclusion the content of the present specification should not manage Solution is limitation of the present invention.

Claims (14)

1. a kind of correcting method of vehicle location error, which is characterized in that the described method includes:
It, will be with the matched winding vision of the winding phenomenon when detecting that winding phenomenon occurs in the process of moving in current vehicle It pose and is merged with the matched vehicle body pose of the winding phenomenon, to generate global fusion map, wherein the winding Vision pose is acquired by the visual sensor of the current vehicle, and the vehicle body pose is adopted by the IMU of the current vehicle Collection;
Based on the key frame in the global fixed winding vision pose of fusion map, and determined to from the key frame Visual signature point optimize, the Target key frames after being optimized;
Winding detection is carried out to the Target key frames, obtains the first winding testing result;
The first winding testing result is fed back into the IMU, is entangled so that the IMU carries out error to the vehicle body pose Just.
2. the method according to claim 1, wherein the method also includes:
When the current vehicle is in driving status, the Current vision pose of the visual sensor acquisition is obtained;
Global optimization is carried out to the Current vision pose, obtains Current vision optimization pose;
It is right when detecting that the Current vision optimization pose indicates that winding phenomenon occurs in the process of moving in the current vehicle The Current vision optimization pose carries out winding detection, obtains the second winding testing result;
The second winding testing result is fed back into the IMU, to correct the running error of the IMU.
3. according to the method described in claim 2, it is characterized in that, it is described to the Current vision pose carry out global optimization, Obtain Current vision optimization pose, comprising:
Pass through the current key frame in the fixed Current vision pose of global optimization;
Optimize the visual signature point in the current key frame, pose is optimized with the Current vision after being optimized.
4. described in any item methods according to claim 1~3, which is characterized in that described to be returned to the Target key frames Ring detection, obtains the first winding testing result, comprising:
It obtains and several matched history keyword frames of the Target key frames;
Calculate the error of the Target key frames Yu each history keyword frame;
Overall error is obtained according to each error calculation, and the overall error is determined as the first winding testing result.
5. according to the method described in claim 4, it is characterized in that, described ought detect that current vehicle occurs in the process of moving When winding phenomenon, will with the matched winding vision pose of the winding phenomenon and with the matched vehicle body pose of the winding phenomenon It is merged, before generating global fusion map, the method also includes:
The environmental visual pose of the current vehicle in the process of moving is acquired by the visual sensor;
When the current environment vision pose and any one of environmental visual pose phase for detecting the visual sensor acquisition Like when, determine that winding phenomenon occurs in the process of moving in the current vehicle.
6. a kind of correcting method of vehicle location error, which is characterized in that the described method includes:
When vehicle is in driving status, the target visual pose of the visual sensor acquisition of the vehicle is obtained;
Pass through the key frame in the fixed target visual pose of global optimization;
Optimize the visual signature point in the key frame, pose is optimized with the target visual after being optimized;
When detecting that the target visual optimization pose indicates that winding phenomenon occurs in the process of moving in the vehicle, to described Target visual optimizes pose and carries out winding detection, obtains third winding result;
The third winding result is fed back to the IMU of the vehicle, to correct the running error of the IMU.
7. a kind of correcting system of vehicle location error characterized by comprising
Associative cell will be with the winding phenomenon for when detecting that winding phenomenon occurs in the process of moving in current vehicle It matched winding vision pose and is merged with the matched vehicle body pose of the winding phenomenon, to generate global fusion ground Figure, wherein the winding vision pose is acquired by the visual sensor of the current vehicle, and the vehicle body pose passes through described The IMU of current vehicle is acquired;
First optimization unit, for fixing the key frame in the winding vision pose based on the global fusion map, and it is right The visual signature point determined from the key frame optimizes, the Target key frames after being optimized;
First detection unit obtains the first winding testing result for carrying out winding detection to the Target key frames;
First feedback unit, for the first winding testing result to be fed back to the IMU, so that the IMU is to the vehicle Body pose carries out error correction.
8. system according to claim 7, which is characterized in that the system also includes:
First acquisition unit, for when the current vehicle is in driving status, obtaining working as the visual sensor acquisition Anterior optic pose;
Second optimization unit obtains Current vision optimization pose for carrying out global optimization to the Current vision pose;
Second detection unit detects that the Current vision optimization pose indicates the current vehicle in the process of moving for working as When there is winding phenomenon, winding detection is carried out to Current vision optimization pose, obtains the second winding testing result;
Second feedback unit, for the second winding testing result to be fed back to the IMU, to correct in the IMU operation Error.
9. system according to claim 8, which is characterized in that described second, which optimizes unit, includes:
Fixed subelement, for passing through the current key frame in the fixed Current vision pose of global optimization;
Optimize subelement, it is excellent with the Current vision after being optimized for optimizing the visual signature point in the current key frame Change pose.
10. according to the described in any item systems of claim 7~9, which is characterized in that the first detection unit includes:
Subelement is obtained, for obtaining and several matched history keyword frames of the Target key frames;
Computation subunit, for calculating the error of the Target key frames Yu each history keyword frame;
It determines subelement, for obtaining overall error according to each error calculation, and the overall error is determined as first time Ring testing result.
11. system according to claim 10, which is characterized in that the system also includes:
Acquisition unit, for the associative cell will with the matched winding vision pose of the winding phenomenon and with described time The matched vehicle body pose of ring phenomenon is merged, and before generating global fusion map, acquires institute by the visual sensor State the environmental visual pose of current vehicle in the process of moving;
Determination unit, for when the current environment vision pose and any one of ring for detecting the visual sensor acquisition When border vision pose is similar, determine that winding phenomenon occurs in the process of moving in the current vehicle.
12. a kind of correcting system of vehicle location error characterized by comprising
Second acquisition unit, the target of the visual sensor acquisition for when vehicle is in driving status, obtaining the vehicle Vision pose;
Fixed cell, for passing through the key frame in the fixed target visual pose of global optimization;
Third optimizes unit, for optimizing the visual signature point in the key frame, is optimized with the target visual after being optimized Pose;
Third detection unit detects that the target visual optimization pose indicates that the vehicle occurs in the process of moving for working as When winding phenomenon, winding detection is carried out to target visual optimization pose, obtains third winding result;
Third feedback unit, for the third winding result to be fed back to the IMU of the vehicle, to correct the IMU operation In error.
13. a kind of vehicle electronic device characterized by comprising
It is stored with the memory of executable program code;
The processor coupled with the memory;
The processor calls the executable program code stored in the memory, and perform claim requires any one of 1~5 The instruction of some or all of the correcting method of vehicle location error step;
Alternatively, the processor calls the executable program code stored in the memory, perform claim is required described in 6 Some or all of the correcting method of vehicle location error step instruction.
14. a kind of vehicle, which is characterized in that the vehicle includes vehicle electronic device described in claim 13.
CN201910421222.3A 2019-05-20 2019-05-20 Method and system for correcting vehicle positioning error Active CN110132306B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910421222.3A CN110132306B (en) 2019-05-20 2019-05-20 Method and system for correcting vehicle positioning error

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910421222.3A CN110132306B (en) 2019-05-20 2019-05-20 Method and system for correcting vehicle positioning error

Publications (2)

Publication Number Publication Date
CN110132306A true CN110132306A (en) 2019-08-16
CN110132306B CN110132306B (en) 2021-02-19

Family

ID=67571804

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910421222.3A Active CN110132306B (en) 2019-05-20 2019-05-20 Method and system for correcting vehicle positioning error

Country Status (1)

Country Link
CN (1) CN110132306B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110967038A (en) * 2019-11-25 2020-04-07 广州小鹏汽车科技有限公司 Vehicle positioning method, vehicle positioning device and vehicle
CN111145634A (en) * 2019-12-31 2020-05-12 深圳市优必选科技股份有限公司 Method and device for correcting map
CN111539305A (en) * 2020-04-20 2020-08-14 肇庆小鹏汽车有限公司 Map construction method and system, vehicle and storage medium
CN112596691A (en) * 2020-09-02 2021-04-02 禾多科技(北京)有限公司 Information pushing method and device, electronic equipment and computer readable medium

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107160395A (en) * 2017-06-07 2017-09-15 中国人民解放军装甲兵工程学院 Map constructing method and robot control system
CN107193279A (en) * 2017-05-09 2017-09-22 复旦大学 Robot localization and map structuring system based on monocular vision and IMU information
US20170343356A1 (en) * 2016-05-25 2017-11-30 Regents Of The University Of Minnesota Resource-aware large-scale cooperative 3d mapping using multiple mobile devices
CN107869989A (en) * 2017-11-06 2018-04-03 东北大学 A kind of localization method and system of the fusion of view-based access control model inertial navigation information
WO2018140701A1 (en) * 2017-01-27 2018-08-02 Kaarta, Inc. Laser scanner with real-time, online ego-motion estimation
CN108447097A (en) * 2018-03-05 2018-08-24 清华-伯克利深圳学院筹备办公室 Depth camera scaling method, device, electronic equipment and storage medium
CN109307508A (en) * 2018-08-29 2019-02-05 中国科学院合肥物质科学研究院 A kind of panorama inertial navigation SLAM method based on more key frames
CN109341706A (en) * 2018-10-17 2019-02-15 张亮 A kind of production method of the multiple features fusion map towards pilotless automobile
CN109579840A (en) * 2018-10-25 2019-04-05 中国科学院上海微系统与信息技术研究所 A kind of close coupling binocular vision inertia SLAM method of dotted line Fusion Features

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170343356A1 (en) * 2016-05-25 2017-11-30 Regents Of The University Of Minnesota Resource-aware large-scale cooperative 3d mapping using multiple mobile devices
WO2018140701A1 (en) * 2017-01-27 2018-08-02 Kaarta, Inc. Laser scanner with real-time, online ego-motion estimation
CN107193279A (en) * 2017-05-09 2017-09-22 复旦大学 Robot localization and map structuring system based on monocular vision and IMU information
CN107160395A (en) * 2017-06-07 2017-09-15 中国人民解放军装甲兵工程学院 Map constructing method and robot control system
CN107869989A (en) * 2017-11-06 2018-04-03 东北大学 A kind of localization method and system of the fusion of view-based access control model inertial navigation information
CN108447097A (en) * 2018-03-05 2018-08-24 清华-伯克利深圳学院筹备办公室 Depth camera scaling method, device, electronic equipment and storage medium
CN109307508A (en) * 2018-08-29 2019-02-05 中国科学院合肥物质科学研究院 A kind of panorama inertial navigation SLAM method based on more key frames
CN109341706A (en) * 2018-10-17 2019-02-15 张亮 A kind of production method of the multiple features fusion map towards pilotless automobile
CN109579840A (en) * 2018-10-25 2019-04-05 中国科学院上海微系统与信息技术研究所 A kind of close coupling binocular vision inertia SLAM method of dotted line Fusion Features

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110967038A (en) * 2019-11-25 2020-04-07 广州小鹏汽车科技有限公司 Vehicle positioning method, vehicle positioning device and vehicle
CN110967038B (en) * 2019-11-25 2021-08-27 广州小鹏自动驾驶科技有限公司 Vehicle positioning method, vehicle positioning device and vehicle
CN111145634A (en) * 2019-12-31 2020-05-12 深圳市优必选科技股份有限公司 Method and device for correcting map
CN111145634B (en) * 2019-12-31 2022-02-22 深圳市优必选科技股份有限公司 Method and device for correcting map
CN111539305A (en) * 2020-04-20 2020-08-14 肇庆小鹏汽车有限公司 Map construction method and system, vehicle and storage medium
CN112596691A (en) * 2020-09-02 2021-04-02 禾多科技(北京)有限公司 Information pushing method and device, electronic equipment and computer readable medium
CN112596691B (en) * 2020-09-02 2021-10-15 禾多科技(北京)有限公司 Information pushing method and device, electronic equipment and computer readable medium

Also Published As

Publication number Publication date
CN110132306B (en) 2021-02-19

Similar Documents

Publication Publication Date Title
CN110132306A (en) The correcting method and system of vehicle location error
KR102296507B1 (en) Method for tracking object by using convolutional neural network including tracking network and computing device using the same
CN110136199A (en) A kind of vehicle location based on camera, the method and apparatus for building figure
CN112734852B (en) Robot mapping method and device and computing equipment
CN109540148A (en) Localization method and system based on SLAM map
CN109084732A (en) Positioning and air navigation aid, device and processing equipment
CN109767475A (en) A kind of method for calibrating external parameters and system of sensor
CN110377025A (en) Sensor aggregation framework for automatic driving vehicle
CN110163889A (en) Method for tracking target, target tracker, target following equipment
CN107990899A (en) A kind of localization method and system based on SLAM
CN106989746A (en) Air navigation aid and guider
CN107796397A (en) A kind of Robot Binocular Vision localization method, device and storage medium
CN109887087A (en) A kind of SLAM of vehicle builds drawing method and system
CN108051002A (en) Transport vehicle space-location method and system based on inertia measurement auxiliary vision
CN105931275A (en) Monocular and IMU fused stable motion tracking method and device based on mobile terminal
CN112219087A (en) Pose prediction method, map construction method, movable platform and storage medium
CN106030430A (en) Multi-sensor fusion for robust autonomous filght in indoor and outdoor environments with a rotorcraft micro-aerial vehicle (MAV)
CN110148170A (en) A kind of positioning initialization method and car-mounted terminal applied to vehicle location
CN110147094A (en) A kind of vehicle positioning method and car-mounted terminal based on vehicle-mounted viewing system
CN109857111A (en) High-precision locating method and system based on shared SLAM map
KR100971778B1 (en) Method, system and computer-readable recording medium for matching of panoramic image using graph structure
CN107478220A (en) Unmanned plane indoor navigation method, device, unmanned plane and storage medium
CN107833249A (en) A kind of carrier-borne aircraft landing mission attitude prediction method of view-based access control model guiding
CN110146096A (en) A kind of vehicle positioning method and device based on image perception
CN109059907A (en) Track data processing method, device, computer equipment and storage medium

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
CB02 Change of applicant information
CB02 Change of applicant information

Address after: 510000 No.8 Songgang street, Cencun, Tianhe District, Guangzhou City, Guangdong Province

Applicant after: GUANGZHOU XPENG AUTOMOBILE TECHNOLOGY Co.,Ltd.

Address before: Room 245, No. 333, jiufo Jianshe Road, Zhongxin Guangzhou Knowledge City, Guangzhou City, Guangdong Province

Applicant before: GUANGZHOU XPENG AUTOMOBILE TECHNOLOGY Co.,Ltd.