CN108873931A - A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness - Google Patents
A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness Download PDFInfo
- Publication number
- CN108873931A CN108873931A CN201810570734.1A CN201810570734A CN108873931A CN 108873931 A CN108873931 A CN 108873931A CN 201810570734 A CN201810570734 A CN 201810570734A CN 108873931 A CN108873931 A CN 108873931A
- Authority
- CN
- China
- Prior art keywords
- image
- unmanned plane
- power line
- distance
- point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000004888 barrier function Effects 0.000 claims abstract description 18
- 238000000034 method Methods 0.000 claims abstract description 18
- 239000011521 glass Substances 0.000 claims abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 12
- RZVHIXYEVGDQDX-UHFFFAOYSA-N 9,10-anthraquinone Chemical compound C1=CC=C2C(=O)C3=CC=CC=C3C(=O)C2=C1 RZVHIXYEVGDQDX-UHFFFAOYSA-N 0.000 claims description 11
- 230000006837 decompression Effects 0.000 claims description 8
- 238000004513 sizing Methods 0.000 claims description 8
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000003384 imaging method Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 230000005611 electricity Effects 0.000 claims description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 4
- 238000012544 monitoring process Methods 0.000 abstract description 3
- 230000000007 visual effect Effects 0.000 abstract description 2
- 230000006835 compression Effects 0.000 description 9
- 238000007906 compression Methods 0.000 description 9
- 238000004364 calculation method Methods 0.000 description 7
- 238000012216 screening Methods 0.000 description 4
- 238000003708 edge detection Methods 0.000 description 3
- 239000000047 product Substances 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 241001347978 Major minor Species 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- HPTJABJPZMULFH-UHFFFAOYSA-N 12-[(Cyclohexylcarbamoyl)amino]dodecanoic acid Chemical compound OC(=O)CCCCCCCCCCCNC(=O)NC1CCCCC1 HPTJABJPZMULFH-UHFFFAOYSA-N 0.000 description 1
- 241000255925 Diptera Species 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 239000010931 gold Substances 0.000 description 1
- 229910052737 gold Inorganic materials 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000004321 preservation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of unmanned plane vision avoiding collisions combined based on subjectiveness and objectiveness, using the method for visual token, image is directly acquired by high definition camera, real-time perfoming target detection, Feature Points Matching scheduling algorithm operation, to obtain range data in real time, unmanned plane can be instructed to make flight processing in real time, keep certain safe distance with target, avoiding barrier, so that safety completes task；Simultaneously, using virtual reality glasses as terminal, the image data that the airborne biocular systems of real-time monitoring obtain allows operator really to perceive surrounding scene by Three-dimensional Display, intervene flight path in time as necessary by unmanned aerial vehicle platform control assembly, further increases safety.
Description
Technical field
The invention belongs to air vehicle technique fields, and in particular to a kind of unmanned plane vision combined based on subjectiveness and objectiveness is anti-
Hit method.
Background technique
Anti-collision system is the unmanned plane safety guarantee important when executing aerial mission, and the reliability of anti-collision system is very big
The level of intelligence of unmanned plane is reflected in degree.For unmanned plane anti-collision system, unmanned plane needs not in flight course
Disconnected real time monitoring ambient enviroment finds barrier in time, and plans that it flies again according to the depth information between barrier
Walking along the street line, to be automatically performed aerial mission.Therefore the most crucial problem of the autonomous anticollision of unmanned plane is solved to obstacle distance
Perception.
Present unmanned plane anti-collision system mainly divides three classes：1) active anti-collision system, this kind of anti-collision system generally can be actively
To barrier project beams, by receiving the feedback information of wave beam to obtain the range information of barrier.Active anticollision system
It unites and perceived distance and anti-barrier is carried out to barrier usually using infrared ray, radar, ultrasonic wave etc..Its measurement model of range-measurement infrared system
It is with limit.Ultrasonic anti-collision system is easy to be influenced by external environment.Its own quality of radar ranging system is bigger than normal, specified function
Rate height and higher cost；2) passive type anti-collision system, this kind of anti-collision system directly acquires the information of scene, then to the figure of acquisition
As carrying out handling and by complicated calculating to obtain the depth information of scene.Passive type anti-collision system it is most typical representative be
Vision anti-collision system, vision anti-collision system simulate the vision system of people, image data are obtained using visual sensor, by figure
Extraneous depth information of scene is obtained as being handled and being calculated, the depth information of usage scenario realizes anticollision, but due to image
The limitation of Processing Algorithm can't reach requirement only according to vision system in some cases；3) composite anti-collision system, it is this kind of
Anti-collision system combines the characteristics of active and passive type avoidance, and system needs multiple sensors, generally include laser ranging,
Binocular Stereo Vision System, IMU etc., multiple sensors improve cost, increase volume and quality, it is difficult to meet small-sized nothing
Man-machine requirement.
Summary of the invention
In view of this, the object of the present invention is to provide a kind of unmanned plane vision anticollision sides combined based on subjectiveness and objectiveness
The reliability of unmanned plane anti-collision system can be improved in method.
A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness, including：Binocular phase is carried on unmanned plane
Machine handles the binocular image of acquisition, obtains the 3-D image of binocular imaging；
Unmanned plane is obtained at a distance from barrier according to the binocular image, and distance information transmission is flown to unmanned plane
Control system, by system for flight control computer automatic obstacle avoiding；
By the load of obstacle distance information on the 3-D image, 3-D image is subjected to video compress, after compression
Video code flow transmitted by wireless data chain, ground real-time reception compression video code flow and decompression, by the figure of decompression
Three-dimensional Display is realized as data are transferred to VR glasses, by eye-observation real time 3-D image and the distance value of superposition, in urgent feelings
It is realized by manual control unmanned plane based on subjective anticollision under condition；
Wherein, after obtaining the binocular image, when barrier is power line, ranging as follows：
The power line in binocular image is detected respectively；For the piece image in binocular image, one is selected on power line
Point P1, according to core collimation method, determining point P1 corresponding same place P2 in the another piece image in binocular image；Further according to point P1 and
Parallax between P2, determines the distance between power line and unmanned plane；
When barrier is the non-electrical line of force：Barrier and unmanned plane are determined using the characteristic point telemetry based on sift algorithm
The distance between, specially：
A, gaussian filtering and down-sampled processing are carried out to binocular image, obtains gaussian pyramid, and then obtain difference gold word
Tower；
B, it is directed to the pyramidal each image of difference, obtains the horizontal gradient Dxx and vertical gradient Dyy of each pixel；
The pixel that the product of horizontal gradient Dxx and vertical gradient Dyy is less than or equal to given threshold is rejected, extreme point scalping is completed
Choosing；
C, for the difference pyramid of completion extreme point coarse sizing, feature point extraction is carried out, the feature of binocular image is obtained
Point；
D, it is directed in binocular image a wherein width, different target is split using watershed algorithm；
E, the characteristic point for obtaining step C gives each target according to the Target Assignment belonging to it；
F, it is directed to each target, selected part characteristic point, and characteristic point is matched, obtains matching characteristic point pair；
Using the parallax of each pair of matching characteristic point, obstacle distance is calculated；
G, multiple obstacle distances corresponding for each target cluster, and obtain the distance of the target；By all mesh
Target minimum range is ultimately sent to flight control system and shows on VR glasses as final obstacle distance.
Preferably, the obstacle distance of every power line is calculated separately, by minimum range when detecting a plurality of power line
It is sent to the flight control system and VR glasses.
Preferably, being less than a plurality of power line of setting value as one to electric power wire spacing when detecting a plurality of power line
Bar power line completes power line cluster；Minimum range is sent to described fly by the obstacle distance for calculating separately every power line
Control system and VR glasses.
The present invention has the advantages that：
1) a kind of, unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness provided by the invention, is surveyed using vision
Away from method, i.e., directly by high definition camera acquire image, real-time perfoming target detection, Feature Points Matching scheduling algorithm operation, from
And range data is obtained in real time, unmanned plane can be instructed to make flight processing in real time, keep certain safe distance with target, evaded
Barrier, so that safety completes task；Meanwhile using virtual reality glasses as terminal, the airborne biocular systems of real-time monitoring are obtained
Image data allows operator really to perceive surrounding scene, as necessary by unmanned aerial vehicle platform control assembly by Three-dimensional Display
Intervene flight path in time, further increases safety.
2), using the specific flying scene of unmanned plane, using power line distance measuring method, i.e., by detection binocular image
Power line, same place is found in binocular image by core collimation method, and determine the parallax of two o'clock, is achieved in ranging, this hair
The bright special flight environment of vehicle using unmanned plane, it is innovative that core collimation method is integrated in barrier ranging, not only opened for unmanned plane
New distance measuring method has been opened up, characteristic point of the same name has been also reduced and carries out the calculating such as matching, reduce calculation amount, improve real-time；In nothing
In the case of power line, the present invention is based on sift matching algorithms to use characteristic point telemetry, but adds before extreme point calculating
New coarse sizing algorithm eliminates a large amount of non-extreme points to reduce calculation amount and further improves real-time.
3), the present invention is handled image using watershed algorithm, entire image is divided into different targets, to same
All characteristic points of one target, selected part characteristic point calculate distance, so as to avoid the repetition of a large amount of same target feature points
It calculates, all distances calculated same target are clustered to obtain the distance of the target, are eventually found all target ranges
The smallest value is simultaneously shown, improves real-time.
Detailed description of the invention
Fig. 1 is the anti-collision system schematic diagram that method of the invention is based on.
Fig. 2 is three-dimensional imaging flow chart of the present invention.
Fig. 3 is power line ranging flow chart of the present invention.
Fig. 4 is feature of present invention point ranging flow chart.
Specific embodiment
The present invention will now be described in detail with reference to the accompanying drawings and examples.
A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness of the invention, technical solution are：At nobody
The real-time acquisition that embedded platform realizes binocular image is carried on machine, and three-dimensional imaging is carried out to the binocular image of acquisition and vision is double
Range estimation is away from the one hand the minimum range information that will acquire is transferred directly to the flight control system of unmanned plane, by system for flight control computer
On the other hand obstacle distance value is superimposed upon on 3-D image by automatic obstacle avoiding, image at this time is carried out H.265 video pressure
Contracting, compressed video code flow is transmitted by wireless data chain, and the video code flow of ground real-time reception compression simultaneously decompresses aobvious
Show, the image data of decompression display is transferred to VR glasses and realizes Three-dimensional Display, by eye-observation real time 3-D image and folds
The distance value added can in case of emergency be realized by manual control based on subjective anticollision.
Referring to Fig.1, what method of the invention used includes binocular camera module based on subjective and objective anti-collision system, embedding
Enter formula processing platform module, wireless sending module, wireless receiving module, decompression display module and virtual reality glasses.Wherein
Embedded processing platform includes mainly the image procossings such as acquisition, compression of images, the object ranging of image；Wireless sending module is
The video image of compression is subjected to network transmission；Wireless receiving module is the video codeword data stream of real-time reception compression；Decompression
Module is that the video code flow that will be received unzips it display；Virtual reality glasses are mainly to carry out to the video image of decompression
Three-dimensional Display.
Referring to Fig. 2, the mainly video image processing on embedded processing platform, its step are as follows.
Step 1: three-dimensional imaging.
(1a) configures two-way COMS chip I MX222, obtains the bayer video data of CMOS output.
The CMOS bayer video data exported is changed into rgb format by (1b).
(1c) carries out 3D correction to the RGB component of two-path video.
RGB after correction is changed into YCbCr422 picture format by (1d)；
(1e) dwindles into every 1920*1080 video all the way the image of 960*1080, and the figure after the diminution of this two-way
As being spliced into 1920*1080 video all the way.
Spliced 1920*1080 video is sent to ARM processing platform by PCIE interface by (1f).
Step 2: using power line ranging, i.e., detecting the electric power in binocular image respectively when barrier is power line
Line；For the piece image in binocular image, a point P1 is selected on power line, according to core collimation method, determines point P1 in binocular figure
Corresponding same place P2 in another piece image as in；Further according to the parallax between point P1 and P2, power line and unmanned plane are determined
The distance between Z；The realization of power line range finder module, referring to Fig. 3, specially：
(2a) after ARM processing platform receives image data, read in image simultaneously be loaded into GPU video memory, by
Ratio edge detection is run on GPU, edge detection is carried out to binocular electric power line image, rolled up using the Ratio template that size is 9*9
Product obtains the binary image comprising marginal information.
(2b) carries out Hough transformation on the basis of binary image, and spatial-domain information is transformed into parameter field and is thrown
Ticket.Screening stage after ballot has the hypothetical constraint of two o'clock：
(2b1) assumes that power line exists in the form of long straight line and runs through entire picture, this is because the binocular in experiment
Vision system possesses lesser field angle.
Between the high-voltage power line of (2b2) list corridor in addition to anti-lightning strike ground wire, remaining is almost parallel trend.Therefore most of electricity
The line of force has angle agreement.
According to above-mentioned two constraint condition, the present invention is using Hough transformation line detection algorithm by background in experiment scene
Interference straight line is well rejected, and may finally detect power line.
(2c) is less than a plurality of power line of setting value as a power line to electric power wire spacing, completes power line cluster；
(2d) selects a point P1 for the piece image in binocular image on power line, according to core collimation method, determines point P1
Corresponding same place P2 in another piece image in binocular image；Further according to the parallax between point P1 and P2, power line is determined
The distance between unmanned plane Z；
Wherein f indicates binocular camera focal length；B indicates the distance between binocular camera optical center；Dx is the width of every pixel in object
Corresponding distance in reason, u1 and u2 are respectively coordinate value of the same place P1 and P2 along u axis, and Z is obstacle distance.
Step 3: when barrier is non-high-voltage power line, using characteristic point telemetry ranging, referring to Fig. 4, specially：
(3a) devises a kind of improvement SIFT feature extraction algorithm based on CUDA for characteristic point ranging, and CPU is received
On the other hand image is stored on the texture memory of GPU by the image of income on the one hand by the memory of image preservation CPU.
(3b) carries out gaussian filtering and down-sampled processing on GPU, to binocular image, obtains gaussian pyramid, and then
Obtain difference pyramid；
Conventional method is to carry out extreme point screening to difference pyramid again below, in the present invention, is carrying out extreme point inspection
Coarse sizing is carried out before surveying, i.e.,：For the pyramidal each image of difference, the horizontal gradient Dxx=of each pixel is obtained | f
(x+1)-f (x-1) | and vertical gradient Dyy=| f (y+1)-f (y-1) |；
Extreme point is the point that pixel changes greatly in image, and DOG operator can generate stronger edge effect, in order to retain
Extreme point in image reduces calculation amount, while unstable marginal point being inhibited to avoid interfering, and carries out pole to difference of Gaussian image
Value point screening design.The Dxx when the point is extreme point, Dyy is bigger, so Dxx*Dyy can obtain a bigger value,
If only one is bigger and another is smaller by mobile rim point then Dxx, Dyy, Dxx*Dyy can obtain one it is smaller
Value, if neither extreme point is also not mobile rim point, and Dxx, Dyy is smaller, Dxx*Dyy can obtain one it is smaller
Value.Therefore, the pixel that the product of horizontal gradient Dxx and vertical gradient Dyy is less than or equal to given threshold is rejected, is passed through
Extreme point coarse sizing is completed in the screening of threshold value T；Wherein T is empirical value.
To the coarse sizing of extreme point, a large amount of non-extreme points are eliminated, the calculating of a large amount of non-extreme points are avoided, to reduce
Calculation amount improves real-time.
For completing the difference pyramid of extreme point coarse sizing, carrying out extreme point calculating and being accurately positioned.
(3c) carries out the auxiliary direction calculating of principal direction to obtained extreme point, on the one hand the major-minor direction of obtained characteristic point is protected
There are on CPU, on the other hand save it on GPU.
(3d) generates feature vector according to the major-minor direction of characteristic point on GPU.
Step 4: using watershed algorithm segmented image.
The width being stored in the binocular image on CPU is converted gray level image by (4a).
(4b) obtains gradient magnitude image with Sobel operator, and it includes two groups of 3*3 matrix operators that Sobel operator, which is a kind of,
Calculation formula is as follows, and wherein A represents original image, DxAnd DyThe image of edge detection is respectively represented,
The horizontal and vertical gradient approximation of each calculated pixel, calculates gradient magnitude by following formula,
Calculation formula is as follows：
Therefore gradient direction is：
(4c) prospect tagged object and calculating, the present invention uses and foreground object and background object is marked respectively, right
Image carries out out operation.
(4d) removes darker spot and limb label.
(4e) carries out supplement to image, carries out supplement to the image of reconstruction.
(4f) superposition prospect is tagged in original image.
(4g) calculates context marker, is split by the method for threshold value to image, is realized by calculating range conversion
Watershed transform crestal line.
The segmentation function of (4h) watershed transform calculates, and image can be made to be equipped with local minimum in certain bits, repaired simultaneously
Change gradient magnitude image, so that having local minimum in prospect and background label pixel；Complete the Target Segmentation in image；
The characteristic point that step 3 obtains is distributed to each target according to its position by (4f)；
(4g) is directed to each target, selected part characteristic point, according to the characteristic point vector calculated in (3d) to characteristic point
It is matched, obtains matching characteristic point pair；Using the parallax of each pair of matching characteristic point, obstacle distance is calculated；Each pair of matching is special
Sign point all obtains an obstacle distance, thus obtains multiple distances；
(4h) multiple obstacle distances corresponding for each target cluster, and obtain the distance of the target；To own
Minimum range in target range is ultimately sent to flight control system and shows on VR glasses as final obstacle distance.
Using watershed segmentation mesh calibration method, unmanned plane observation ability is enhanced, it can be very intuitive anti-by image
Reflect the range information of specific objective；Carry out a large amount of characteristic points due to will detect that in scene, when some characteristic points belong to the same mesh
When mark, selected part characteristic point ranging does not need all characteristic points of a target all to carry out ranging operation, reduces
Calculation amount, improves real-time.
Step 5: H.265 video compress and transmission：
In the case where guaranteeing identical image quality, half H.265 is reduced than H.264 video transmission bandwidth.Make on ARM
With GStreamer, by treated, H.265 video image is compressed, and code rate 4Mbps passes through the video code flow of compression
Wireless transport module carries out network transmission.
Step 6: Three-dimensional Display：
The video code flow of compression is received by wireless receiving module in earth station, and video code flow is passed through into software on ground solution
Compression display is transmitted to VR virtual reality glasses by HDMI interface, so as to intuitively observe distance value and by manual control
Prevent unmanned plane anticollision.On the other hand the smallest distance value can be sent to system for flight control computer, is controlled by flight control system
Unmanned plane is to realize unmanned plane anticollision.
In conclusion the above is merely preferred embodiments of the present invention, being not intended to limit the scope of the present invention.
All within the spirits and principles of the present invention, any modification, equivalent replacement, improvement and so on should be included in of the invention
Within protection scope.
Claims (3)
1. a kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness, which is characterized in that including：It is taken on unmanned plane
Binocular camera is carried, the binocular image of acquisition is handled, the 3-D image of binocular imaging is obtained；
Unmanned plane is obtained at a distance from barrier according to the binocular image, and by distance information transmission to the winged control system of unmanned plane
System, by system for flight control computer automatic obstacle avoiding；
By the load of obstacle distance information on the 3-D image, 3-D image is subjected to video compress, by compressed view
Frequency code stream is transmitted by wireless data chain, the video code flow that ground real-time reception compresses and decompression, by the picture number of decompression
Three-dimensional Display is realized according to VR glasses are transferred to, by eye-observation real time 3-D image and the distance value of superposition, in case of emergency
It is realized by manual control unmanned plane based on subjective anticollision；
Wherein, after obtaining the binocular image, when barrier is power line, ranging as follows：
The power line in binocular image is detected respectively；For the piece image in binocular image, a point P1 is selected on power line,
According to core collimation method, determining point P1 corresponding same place P2 in the another piece image in binocular image；Further according to point P1 and P2 it
Between parallax, determine the distance between power line and unmanned plane；
When barrier is the non-electrical line of force：It is determined between barrier and unmanned plane using the characteristic point telemetry based on sift algorithm
Distance, specially：
A, gaussian filtering and down-sampled processing are carried out to binocular image, obtains gaussian pyramid, and then obtain difference pyramid；
B, it is directed to the pyramidal each image of difference, obtains the horizontal gradient Dxx and vertical gradient Dyy of each pixel；By water
The pixel that the product of flat ladder degree Dxx and vertical gradient Dyy is less than or equal to given threshold is rejected, and extreme point coarse sizing is completed；
C, for the difference pyramid of completion extreme point coarse sizing, feature point extraction is carried out, the characteristic point of binocular image is obtained；
D, it is directed in binocular image a wherein width, different target is split using watershed algorithm；
E, the characteristic point for obtaining step C gives each target according to the Target Assignment belonging to it；
F, it is directed to each target, selected part characteristic point, and characteristic point is matched, obtains matching characteristic point pair；It utilizes
The parallax of each pair of matching characteristic point calculates obstacle distance；
G, multiple obstacle distances corresponding for each target cluster, and obtain the distance of the target；By all targets
Minimum range is ultimately sent to flight control system and shows on VR glasses as final obstacle distance.
2. a kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness as described in claim 1, which is characterized in that
When detecting a plurality of power line, the obstacle distance of every power line is calculated separately, minimum range is sent to the winged control
System and VR glasses.
3. a kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness as described in claim 1, which is characterized in that
When detecting a plurality of power line, a plurality of power line of setting value is less than as a power line to electric power wire spacing, completes electricity
Line of force cluster；Minimum range is sent to the flight control system and VR glasses by the obstacle distance for calculating separately every power line.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810570734.1A CN108873931A (en) | 2018-06-05 | 2018-06-05 | A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810570734.1A CN108873931A (en) | 2018-06-05 | 2018-06-05 | A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108873931A true CN108873931A (en) | 2018-11-23 |
Family
ID=64336755
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810570734.1A Pending CN108873931A (en) | 2018-06-05 | 2018-06-05 | A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108873931A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110244760A (en) * | 2019-06-06 | 2019-09-17 | 深圳市道通智能航空技术有限公司 | A kind of barrier-avoiding method, device and electronic equipment |
| CN112148033A (en) * | 2020-10-22 | 2020-12-29 | 广州极飞科技有限公司 | Method, device and equipment for determining unmanned aerial vehicle air route and storage medium |
| CN114115278A (en) * | 2021-11-26 | 2022-03-01 | 东北林业大学 | Obstacle avoidance system based on FPGA (field programmable Gate array) for forest fire prevention robot during traveling |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0459295A2 (en) * | 1990-05-25 | 1991-12-04 | Toshiba Electronic Systems Co., Ltd. | Aircraft docking guidance system which takes position reference in anti-collision light of aircraft |
| EP2130766A2 (en) * | 2008-06-04 | 2009-12-09 | Honeywell International Inc. | Anti-collision lighting systems and methods for a micro aerial vehicle |
| CN101763119A (en) * | 2009-12-16 | 2010-06-30 | 东南大学 | Obstacle avoidance aiding method based on teleoperation mobile robot |
| CN106340009A (en) * | 2016-08-18 | 2017-01-18 | 河海大学常州校区 | Parallel-binocular-based power line detection method and system |
| CN106681353A (en) * | 2016-11-29 | 2017-05-17 | 南京航空航天大学 | Unmanned aerial vehicle (UAV) obstacle avoidance method and system based on binocular vision and optical flow fusion |
| CN106708074A (en) * | 2016-12-06 | 2017-05-24 | 深圳市元征科技股份有限公司 | Method and device for controlling unmanned aerial vehicle based on VR glasses |
| CN107462217A (en) * | 2017-07-07 | 2017-12-12 | 北京航空航天大学 | Unmanned aerial vehicle binocular vision barrier sensing method for power inspection task |
| CN108037765A (en) * | 2017-12-04 | 2018-05-15 | 国网山东省电力公司电力科学研究院 | A kind of unmanned plane obstacle avoidance system for polling transmission line |
-
2018
- 2018-06-05 CN CN201810570734.1A patent/CN108873931A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0459295A2 (en) * | 1990-05-25 | 1991-12-04 | Toshiba Electronic Systems Co., Ltd. | Aircraft docking guidance system which takes position reference in anti-collision light of aircraft |
| EP2130766A2 (en) * | 2008-06-04 | 2009-12-09 | Honeywell International Inc. | Anti-collision lighting systems and methods for a micro aerial vehicle |
| CN101763119A (en) * | 2009-12-16 | 2010-06-30 | 东南大学 | Obstacle avoidance aiding method based on teleoperation mobile robot |
| CN106340009A (en) * | 2016-08-18 | 2017-01-18 | 河海大学常州校区 | Parallel-binocular-based power line detection method and system |
| CN106681353A (en) * | 2016-11-29 | 2017-05-17 | 南京航空航天大学 | Unmanned aerial vehicle (UAV) obstacle avoidance method and system based on binocular vision and optical flow fusion |
| CN106708074A (en) * | 2016-12-06 | 2017-05-24 | 深圳市元征科技股份有限公司 | Method and device for controlling unmanned aerial vehicle based on VR glasses |
| CN107462217A (en) * | 2017-07-07 | 2017-12-12 | 北京航空航天大学 | Unmanned aerial vehicle binocular vision barrier sensing method for power inspection task |
| CN108037765A (en) * | 2017-12-04 | 2018-05-15 | 国网山东省电力公司电力科学研究院 | A kind of unmanned plane obstacle avoidance system for polling transmission line |
Non-Patent Citations (3)
| Title |
|---|
| GULIANG89: "标记分水岭分割算法", 《百度文库》 * |
| 周士超: "无人机电力巡检视觉避撞技术研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
| 牛明花 等: "基于机器视觉的车道线识别算法", 《天津职业技术师范大学学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110244760A (en) * | 2019-06-06 | 2019-09-17 | 深圳市道通智能航空技术有限公司 | A kind of barrier-avoiding method, device and electronic equipment |
| CN112148033A (en) * | 2020-10-22 | 2020-12-29 | 广州极飞科技有限公司 | Method, device and equipment for determining unmanned aerial vehicle air route and storage medium |
| CN114115278A (en) * | 2021-11-26 | 2022-03-01 | 东北林业大学 | Obstacle avoidance system based on FPGA (field programmable Gate array) for forest fire prevention robot during traveling |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10678257B2 (en) | Generating occlusion-aware bird eye view representations of complex road scenes | |
| US20230014874A1 (en) | Obstacle detection method and apparatus, computer device, and storage medium | |
| US11106203B2 (en) | Systems and methods for augmented stereoscopic display | |
| CN113359810B (en) | Unmanned aerial vehicle landing area identification method based on multiple sensors | |
| Vandapel et al. | Natural terrain classification using 3-d ladar data | |
| US20210150227A1 (en) | Geometry-aware instance segmentation in stereo image capture processes | |
| US9269145B2 (en) | System and method for automatically registering an image to a three-dimensional point set | |
| US9275267B2 (en) | System and method for automatic registration of 3D data with electro-optical imagery via photogrammetric bundle adjustment | |
| CN111527463A (en) | Method and system for multi-target tracking | |
| US20100305857A1 (en) | Method and System for Visual Collision Detection and Estimation | |
| CN112740269A (en) | Target detection method and device | |
| Rong et al. | Intelligent detection of vegetation encroachment of power lines with advanced stereovision | |
| CN108873931A (en) | A kind of unmanned plane vision avoiding collision combined based on subjectiveness and objectiveness | |
| CN109213138B (en) | Obstacle avoidance method, device and system | |
| CN111983603A (en) | Motion trajectory relay method, system and device and central processing equipment | |
| CN114611635B (en) | Object identification method and device, storage medium and electronic device | |
| CN114862952B (en) | Unmanned aerial vehicle detection and defense method and system | |
| Kim et al. | Fusing lidar data and aerial imagery with perspective correction for precise localization in urban canyons | |
| JP2022511147A (en) | Systems and methods to facilitate the generation of geographic information | |
| KR102588386B1 (en) | Method and apparatus for detecting obscured object using a lidar | |
| Shimoni et al. | Detection of vehicles in shadow areas | |
| CN117423077A (en) | BEV perception model, construction method, device, equipment, vehicle and storage medium | |
| Baeck et al. | Drone based near real-time human detection with geographic localization | |
| CN113836975A (en) | Binocular vision unmanned aerial vehicle obstacle avoidance method based on YOLOV3 | |
| Su | Vanishing points in road recognition: A review |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20181123 |
|
| WD01 | Invention patent application deemed withdrawn after publication |