CN103386975A - Vehicle obstacle avoidance method and system based on machine vision - Google Patents
Vehicle obstacle avoidance method and system based on machine vision Download PDFInfo
- Publication number
- CN103386975A CN103386975A CN2013103337811A CN201310333781A CN103386975A CN 103386975 A CN103386975 A CN 103386975A CN 2013103337811 A CN2013103337811 A CN 2013103337811A CN 201310333781 A CN201310333781 A CN 201310333781A CN 103386975 A CN103386975 A CN 103386975A
- Authority
- CN
- China
- Prior art keywords
- obstacle
- vehicle
- image
- camera
- controller
- 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
Links
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention provides a vehicle obstacle avoidance method and a vehicle obstacle avoidance system based on machine vision. The obstacle avoidance method comprises the following steps: judging whether an obstacle exists in front of a micro vehicle by a processor by using a obstacle detection algorithm based on a single image; calibrating two cameras, determining the height of the obstacle through a stereoscopic vision method and transmitting height information to a controller by the processor; detecting conditions of vehicles in two adjacent lanes by distance measuring equipment to provide a running area for the vehicles to avoid the obstacle and change a lane and transmitting running area information to the controller; and according to the obtained information, transmitting a command to a operation and control module by the controller through using an adaptive lane change strategy to complete autonomous lane change of the vehicles. The obstacle avoidance method provided by the invention has the advantages of being stable, high in adaptive degree, fluent in vehicle obstacle avoidance posture and up to 98% in obstacle avoidance success rate.
Description
Technical field
The present invention relates to the intelligent transport technology field, be specifically related to a kind of method and system of vehicle obstacle-avoidance based on machine vision.
Background technology
Along with the quickening of urbanization process, the development that existing traffic infrastructure and management method fall behind the times.Merely by widening road, build overhead, lay track traffic, set up sign, encourage to take public transport, even air communication, relies on traditional method can not adapt to the development of Modern Traffic far away, need to develop intelligent transportation system.Intelligent transportation system be with advanced person's information techenology, data communication transmission, control technology and artificial intelligence technology etc. effectively integrated use set up in whole transport administration system on a large scale, comprehensive play a role in real time, accurately, the comprehensive traffic commander of efficient transportation, management and control system.
Intelligent transportation system results from the end of the sixties in last century.Since the eighties in last century, the research in this field has entered the stage of a develop rapidly.The west Main Economic poweies such as the U.S., Japan, Canada, Germany, France all drop into a large amount of manpower and materials to this, can say that " intelligent transportation " is the important symbol that communications and transportation enters the information age.Advanced person's intelligent transportation system is applied to existing traffic facilities, can effectively reduce traffic loading and environmental pollution, assurance safety traffic, improve conveying efficiency, promote socio-economic development, improve people's living standard, and can promote the formation of social informatization and new industry.The more important thing is,, along with constantly advancing of state-of-the-art technology, also make traffic intelligence that the possibility that realizes has been arranged.As one of trend of future transportation development, the relevant departments such as the Chinese government and science and technology, transport administration pay much attention to and actively promote the development of intelligent transportation system.Intelligent transportation system will form a huge market in the development and application of China, and scale can be 10,000,000,000 even more than hundred billion yuan, and it certainly will produce huge promotion to various aspects such as the road of China, traffic, communication, telecommunications, transport administrations.
As one of important research content of intelligent transportation system, intelligent vehicle is driven main research driverless operation technology or as DAS (Driver Assistant System), is helped chaufeur to complete the vehicular drive task.These tasks comprise the tracking road, keep Vehicle Driving Cycle on correct road, keep a safety distance between vehicle,, according to the speed of current traffic and roadway characteristic adjusting vehicle, across track, to reach, overtake other vehicles and keep away the purpose of barrier and find the shortest path that reaches destination and travel easily in urban district and stop.Intelligent vehicle control loop concentrated area has used the technology such as computing machine, sensor, information fusion, communication, artificial intelligence and automatic control, is typical new and high technology synthesis.The intelligent vehicle control loop will effectively alleviate the burden of chaufeur, reduce the phenomenon of driver tired driving, be conducive to improve safety traffic, simultaneously, coordinate urban traffic control system, and the reasonable distribution flow of traffic, realize smooth traffic.Along with the develop rapidly of computing machine and Robotics, intelligent vehicle research makes significant progress, and is widely used in military affairs, scientific research, the every field such as civilian.At military aspect, intelligent vehicle can replace the soldier to complete the tasks such as battle reconnaissance in danger zone; Aspect scientific research, intelligent vehicle can be engaged in exploration and wait work extraterrestrial; Aspect civilian, can be used as automatically or DAS (Driver Assistant System) reduces traffic accident.
Vision system mainly plays the effect of environment detection and identification in Intelligent Vehicle System.Compare with other sensor, computer vision have detect contain much information, can remote measurement etc. advantage.Shortcoming is under complex environment, the target of surveying be separated with background area, useful information is extracted required calculated amount very large, solves with hardware condition merely, easily causes the real-time of system poor.In present the intelligent vehicles technology, autonomous navigation and automatic Pilot are the gordian techniquies of intelligent vehicle exploitation, and in the implementation procedure of autonomous navigation and automatic Pilot, very crucial technology has been the identification of obstacle and kept away barrier, and this is a technical matters of needing solution badly.
Summary of the invention
In order to overcome the defect that exists in above-mentioned prior art, the purpose of this invention is to provide a kind of method and system of vehicle obstacle-avoidance based on machine vision, the stability of keeping away barrier success ratio, Vehicle Driving Cycle that improves in the Vehicle Driving Cycle process is stronger.
In order to realize above-mentioned purpose of the present invention, according to an aspect of the present invention, the invention provides a kind of method of vehicle obstacle-avoidance based on machine vision, comprise the steps:
S1: processor adopting judges based on the detection of obstacles algorithm of single image whether miniature front side exists obstacle, if obstacle is arranged, with the detection communication in obstacle distance and obstacle track of living in to controller, and execution step S2,, if there is no obstacle, continue execution step S1;
S2: treater is demarcated two cameras, adopts the method for stereovision to judge the three-dimensional coordinate of obstacle center, center, coboundary, thus determine the height of obstacle and with described obstacle height communication to controller;
S3: the distance measuring equipment (DME) of vehicle body both sides detects the vehicle condition in adjacent two tracks, for the Bi Zhanghuan road provides the zone of wheeled and with described, exercises regional communication to described controller;
S4: controller is according to the exercised area information in obstacle distance, the obstacle track of living in, obstacle height and adjacent two tracks that obtain, adopt self adaptation to change strategy, send order to operation control module, control turning to and speed of wheel, complete independently changing of vehicle.
In a kind of preferred implementation of the present invention, the step of the detection of obstacles algorithm of described single image is:
S21: image is carried out colour space transformation RGB turn HSV, and extract separately the S channel image;
S22: the large Tianjin of S imagery exploitation method is carried out the dynamic threshold binaryzation obtain bianry image Bin,, to the Bin image denoising, eliminate and disturb;
S23: the obstacle profile that search institute likely exists on the Bin image, the obstacle rejecting that will judge by accident according to the screening conditions that arrange;
S24: and according to the direction deciding obstacle track of living in of lane mark.
In a kind of preferred implementation of the present invention, the method for described stereovision is:
S31:, according to the installation site of main camera A and auxiliary camera B, demarcate respectively, draw the intrinsic parameter of camera, outer parameter, and the spacing b between main camera A and auxiliary camera B;
S32: the image that auxiliary camera B is obtained adopts the obstacle detection method based on single image, obtains obstacle center, the two-dimensional coordinate of center, coboundary under image coordinate system;
S33: utilize following formula to obtain the three-dimensional coordinate x of obstacle under main camera system of axes
c, y
c, z
c, and utilize the calibrating parameters of camera, draw the three-dimensional coordinate of obstacle under world coordinate system,
Wherein, b is the spacing of two cameras, and f is the focal length (u of camera
1, v
1) be the projected position of calibration point on A camera plane, (u
2, v
2) be the projected position of calibration point on B camera plane, and v
1=v
2
In another kind of preferred implementation of the present invention, described sectional type ADAPTIVE CONTROL refers to: deviation angle and the distance middle apart from track according to vehicle body, and vehicle body moving velocity this moment, adjust next vehicle body corner and speed of a motor vehicle constantly according to parameter d egree_per_dist self adaptation appropriateness, when near the centre of Vehicle Driving Cycle at straight way during position, remain the direction running that vehicle body is kept straight at full speed, if any side of deflection both sides, track is finely tuned.
In order to realize above-mentioned purpose of the present invention, according to another aspect of the present invention, the invention provides a kind of system of vehicle obstacle-avoidance based on machine vision, comprise main camera A, auxiliary camera B, distance measuring equipment (DME), treater and controller, described main camera A and auxiliary camera B be respectively used to obtain the image on road surface and with described image transmitting to described treater; Described treater be used for judging whether miniature front side exists obstacle and with the detection communication in obstacle distance and obstacle track of living in to controller, simultaneous processor is demarcated main camera A and auxiliary camera B, judge the three-dimensional coordinate of obstacle center, center, coboundary, thus determine the height of obstacle and with described obstacle height communication to controller; Described distance measuring equipment (DME) is positioned at the vehicle body both sides, for detection of the vehicle condition in adjacent two tracks, also can exercise the communication in zone to described controller; Controller is connected with operation control module, described operation control module is connected with wheel, controller is according to the exercised area information in obstacle distance, the obstacle track of living in, obstacle height and adjacent two tracks that obtain, send order to operation control module, described operation control module is controlled turning to and speed of wheel, completes independently changing of vehicle.
The invention has the beneficial effects as follows:
At first the present invention adopts based on the detection of obstacles algorithm of single image to judge whether the place ahead exists obstacle, the inadequate problem of real-time of having avoided direct employing stereo visual system to detect causing, and sense cycle is short; Then further utilize dual camera, use stereo visual system to judge the elevation information of obstacle; Barrier-avoiding method of the present invention is stable, the self adaptation degree is high, and the vehicle obstacle-avoidance attitude is smooth, keeps away the barrier success ratio up to more than 98%.
The present invention adopts the elevation information of the method obstacle of stereovision, has taken into account real-time and the accuracy requirement of system.
The present invention's state model residing according to vehicle, adopt the sectional type ADAPTIVE CONTROL, and bottom controller is sent control command.Deviation angle and the distance middle apart from track according to vehicle body, and vehicle body moving velocity this moment, adjust next vehicle body corner and speed of a motor vehicle constantly according to parameter d egree_per_dist self adaptation appropriateness.Straight way is as the special road type of bend,, when near the centre of Vehicle Driving Cycle at straight way during position, remains the direction running that vehicle body is kept straight at full speed, if any side of deflection both sides, track is carried out appropriate amount of fine-tuning according to the sectional type ADAPTIVE CONTROL.Other control methods of comparing, in the present invention, method can effectively reduce the adjustment number of times of vehicle body steering wheel angle, break away from vehicle body and adjust because of frequent the vehicle body wigwag motion that travel direction causes, thereby ensureing that vehicle body is metastable travels.
Sectional type ADAPTIVE CONTROL of the present invention can effectively reduce the adjustment number of times of vehicle body steering wheel angle, break away from vehicle body and adjust because of frequent the vehicle body wigwag motion that travel direction causes, thereby ensureing that vehicle body is metastable travels.
Additional aspect of the present invention and advantage part in the following description provide, and part will become obviously from the following description, or by practice of the present invention, recognize.
Description of drawings
Above-mentioned and/or additional aspect of the present invention and advantage are from obviously and easily understanding becoming the description of embodiment in conjunction with following accompanying drawing, wherein:
Fig. 1 is miniature intelligent vehicle hardware structure in a kind of preferred implementation of the present invention;
Fig. 2 is the vehicle obstacle-avoidance method flow diagram that the present invention is based on machine vision.
The specific embodiment
Below describe embodiments of the invention in detail, the example of described embodiment is shown in the drawings, and wherein same or similar label represents same or similar element or the element with identical or similar functions from start to finish.Be exemplary below by the embodiment that is described with reference to the drawings, only be used for explaining the present invention, and can not be interpreted as limitation of the present invention.
In description of the invention, unless otherwise prescribed and limit, need to prove, term " installation ", " being connected ", " connection " should be done broad understanding, for example, can be mechanical connection or electrical connection, can be also the connection of two element internals, can be directly to be connected, and also can indirectly be connected by intermediary, for the ordinary skill in the art, can understand as the case may be the concrete meaning of above-mentioned term.
The present invention can be applied to actual vehicle, also can be applied to miniature car, in a kind of preferred implementation of the present invention, adopts miniature car to describe, and the concrete miniature car that adopts obtains with the miniature ratio of true car according to 1:10.
The invention provides a kind of miniature car based on machine vision, it comprises chassis, is provided with motor and steering wheel on this chassis.This miniature car also comprises camera, mainboard and control desk, and camera is used for obtaining the road surface picture, and the road surface picture is used for the information such as identification lane mark, road markings and obstacle; The treater of mainboard is connected with camera, mainboard moves control program and assigns control command according to the information of camera transmission, control desk is connected with mainboard, and the control command that control desk reception mainboard is assigned is controlled the operation of motor and steering wheel by operation control module, control advancing of wheel.
The present invention also provides a kind of miniature intelligent vehicle obstacle avoidance system based on machine vision, as shown in Figure 1, comprise main camera A, auxiliary camera B, distance measuring equipment (DME), treater and controller, described main camera A and auxiliary camera B be respectively used to obtain the image on road surface and with described image transmitting to described treater; Described treater be used for judging whether miniature front side exists obstacle and with the detection communication in obstacle distance and obstacle track of living in to controller, simultaneous processor is demarcated main camera A and auxiliary camera B, judge the three-dimensional coordinate of obstacle center, center, coboundary, thus determine the height of obstacle and with described obstacle height communication to controller; Described distance measuring equipment (DME) is positioned at the vehicle body both sides, for detection of the vehicle condition in adjacent two tracks, also can exercise the communication in zone to described controller; Controller is connected with operation control module, described operation control module is connected with wheel, controller is according to the exercised area information in obstacle distance, the obstacle track of living in, obstacle height and adjacent two tracks that obtain, send order to operation control module, described operation control module is controlled turning to and speed of wheel, completes independently changing of vehicle.
In the present embodiment, this miniature chassis is selected HPI cup racer chassis, and this chassis, with motor and steering wheel, is used for mechanical movement.Control desk adopts the Arduino control desk of DFR0003 model.Mainboard adopts embedded x86 mainboard.Camera can be 1, also can be for a plurality of, when camera while being a plurality of, each camera all is connected with mainboard, and in a preferred enforcement side of the present invention, camera is two, in one of the present invention was more preferably enforcement side, camera was sieve skill camera.
The road surface pictorial information that miniature car based on machine vision of the present invention obtains according to vision sensor by mainboard and control desk is controlled the operation of motor and steering wheel, this miniature car is with respect to the archeus vehicle, miniature intelligent vehicle simple in structure, cheap, many cars test environment easily builds.And experimental site and environment are easily adjusted, and can carry out easily the experiment under multiple varying environment.
The hardware structure of the miniature car of the present invention comprises the car load hardware structure that is complementary with bodywork length, space structure reasonable in design, in the present embodiment, the vehicle body space of miniature car is divided into the upper, middle and lower-ranking structure, places corresponding hardware unit in every layer of structure.The hardware such as control desk, chassis is positioned at vehicle body lower floor, places mainboard in middle level, at upper strata vehicle body spatial placement vision sensor.Bottom controller and treater are powered respectively.Image-pickup device is installed on the top of intelligent vehicle apart from 20-25cm place, ground.
In a kind of preferred implementation of the present invention, this miniature car also comprises for the electronic compass of judgement service direction and uphill/downhill angle with for the infrared distance sensor that judges front-and-rear vehicle distance and the miniature spacing of adjacent lane, electronic compass is connected with control desk respectively with infrared distance sensor, and control desk is controlled the operation of motor and steering wheel according to the information of electronic compass and infrared distance sensor transmission.
In a preferred embodiment of the present invention, infrared distance sensor is two-way, and a road is used for the judgement front-and-rear vehicle distance, and another road is used for the distance of the miniature car of this miniature car of judgement and adjacent lane.
The present invention, by utilizing electronic compass and infrared distance sensor, has realized the accurate control to miniature Vehicle Driving Cycle, has greatly improved the safety of travelling.
In the present embodiment, this miniature car also comprises power supply, in a kind of preferred implementation of the present invention, adopts the 12V lithium cell to main board power supply; Adopt the 8V lithium cell to power to motor; Vision sensor and control desk are by main board power supply.
In the present embodiment, miniature car also comprises the car shell, for the ease of realizing the research to miniature car, need to consider the arrangement situation of each accessory, comprise weight, size factor, due to the X86 mainboard and the motor consumption of current high, if select high power battery can increase miniature car weight amount, therefore need preferred dimension little, the battery that power is suitable, selecting of car shell also needs to consider size factor simultaneously.In a kind of preferred implementation of the present invention, follow following principle during configuration design:
1, mainboard battery and motor battery can install easily, dismantle, to facilitate charging;
2, the control desk battery need not frequent dismounting, but must reserve charge port;
3, the car shell must be for convenience detach, and battery changes the outfit.Can consider battery and car shell Uniting according to the big or small customized vehicle shell of vehicle body, be convenient for changing battery.
The present invention also provides a kind of miniature intelligent vehicle barrier-avoiding method based on machine vision, and as shown in Figure 2, it comprises the steps:
S1: processor adopting judges based on the detection of obstacles algorithm of single image whether miniature front side exists obstacle, if obstacle is arranged, with the detection communication in obstacle distance and obstacle track of living in to controller, and execution step S2,, if there is no obstacle, continue execution step S1;
S2: treater is demarcated two cameras, adopts the method for stereovision to judge the three-dimensional coordinate of obstacle center, center, coboundary, thus determine the height of obstacle and with described obstacle height communication to controller;
S3: the distance measuring equipment (DME) of vehicle body both sides detects the vehicle condition in adjacent two tracks, for the Bi Zhanghuan road provides the zone of wheeled and with described, exercises regional communication to described controller;
S4: controller is according to the exercised area information in obstacle distance, the obstacle track of living in, obstacle height and adjacent two tracks that obtain, adopt self adaptation to change strategy, send order to operation control module, control turning to and speed of wheel, complete independently changing of vehicle.
In the present embodiment, the step of the detection of obstacles algorithm of described single image is:
S21: image is carried out colour space transformation RGB turn HSV, and extract separately the S channel image;
S22: the large Tianjin of S imagery exploitation method is carried out the dynamic threshold binaryzation obtain bianry image Bin,, to the Bin image denoising, eliminate and disturb;
S23: the obstacle profile that search institute likely exists on the Bin image, the obstacle rejecting that will judge by accident according to the screening conditions that arrange;
S24: and according to the direction deciding obstacle track of living in of lane mark.
In the present embodiment, the method for described stereovision is:
S31: according to the installation site of main camera A and auxiliary camera B, demarcate respectively, draw the intrinsic parameter of camera, outer parameter, and the spacing b between main camera A and auxiliary camera B, in the present embodiment, the intrinsic parameter of camera, outer parameter can draw according to acquisition methods of the prior art;
S32: the image that auxiliary camera B is obtained adopts the obstacle detection method based on single image, obtains obstacle center, the two-dimensional coordinate of center, coboundary under image coordinate system;
S33: utilize following formula to obtain the three-dimensional coordinate x of obstacle under main camera system of axes
c, y
c, z
c, and utilize intrinsic parameter and the outer parameter of the demarcation of camera, draw the three-dimensional coordinate of obstacle under world coordinate system,
Wherein, b is the spacing of two cameras, and f is the focal length of camera,, (u
1, v
1) be the projected position of calibration point on A camera plane, (u
2, v
2) be the projected position of calibration point on B camera plane, in the present embodiment, calibration point equates that with the ordinate of projected position on B camera plane v is namely arranged at the projected position on A camera plane
1=v
2
In the present embodiment, described sectional type ADAPTIVE CONTROL refers to: deviation angle and the distance middle apart from track according to vehicle body, and vehicle body moving velocity this moment, adjust next vehicle body corner and speed of a motor vehicle constantly according to parameter d egree_per_dist self adaptation appropriateness, when near the centre of Vehicle Driving Cycle at straight way during position, remain the direction running that vehicle body is kept straight at full speed, if any side of deflection both sides, track is finely tuned.
The present invention also had following steps before step S1: carry out line walking by inspection system, determine lane mark.In the present embodiment, the inspection system of miniature car comprises image collection module, mainboard and controller, described image collection module be used for obtaining the vision sensor collection road surface the RGB coloured image and described RGB coloured image is transferred to described mainboard, contain lane mark information in described RGB coloured image; Described mainboard comprises the line walking module, and described line walking module comprises image processing module, binarization block, rim detection module and treater; The RGB coloured image that the described image collection module of described image processing module reception is obtained also is converted into gray level image with described RGB coloured image; Described binarization block is connected with described image processing module, is used for receiving described gray level image and obtaining the optimal dynamic threshold value of the every two field picture of described gray level image and carry out image segmentation, obtains bianry image, and lane mark is separated; Described rim detection module is connected with described binarization block, be used for to receive and described binary image is carried out rim detection, obtains containing the edge image at the inside and outside edge of lane mark; Described treater is connected with described rim detection module, and described treater utilizes Hough transformation inspection vehicle diatom and obtain the lane mark parameter in described edge image, sets up the lane mark model; The described lane mark parameter that the treater utilization is obtained, obtain vehicle residing lane position data in world coordinate system by inverse perspective mapping, and vehicle corner and distance parameter, differentiates the driving mode of vehicle; Utilize simultaneously vehicle corner and the distance parameter that obtains,, according to the driving mode of vehicle, adopt the sectional type ADAPTIVE CONTROL, to controller, send control command; Described controller is connected with treater, steering wheel and motor respectively, is used for receiving the control command of described treater, and adjusts the working parameter of steering wheel and motor according to control command, controls in real time direction of traffic and the road speed of vehicle.
In the present embodiment, mainboard also comprises initialization module, detection of obstacles module, traffic lights detection module, traffic sign detection module, surface mark detection module; Described initialization module is connected with described vision sensor, described motor and described steering wheel respectively, is used for described vision sensor, described motor and described steering wheel are carried out initialization; Described detection of obstacles module is connected with described vision sensor, is used for the obstacle of the road surface image detect vehicle front that obtains according to described vision sensor; Described traffic lights detection module is connected with described vision sensor, is used for the service condition of the road surface image detect traffic lights that obtain according to described vision sensor; Described traffic sign detection module is connected with described vision sensor, is used for traffic sign being identified and being judged according to the road surface picture that described vision sensor obtains; Described surface mark detection module is connected with described vision sensor, is used for the surface mark in the picture recognition vehicle track of living in, road surface that obtains according to described vision sensor.The present invention carries out initialization by the initialization module of mainboard to vision sensor, motor and steering wheel, improved the accuracy of travelling, in addition, the hunting module that this mainboard can have, detection of obstacles module, traffic lights detection module, traffic sign detection module, surface mark detection module, realized that automated lane line following, site of road detect, certainly move, many cars are interactive and analyzed the ability of vehicle target direction, improved the safety of travelling.
In another kind of preferred implementation of the present invention, mainboard also comprises rate control module and direction control module, and rate control module is connected with initialization module, hunting module, detection of obstacles module, traffic lights detection module, traffic sign detection module and surface mark detection module respectively with described direction control module.Rate control module receives the information of described initialization module, hunting module, detection of obstacles module, traffic lights detection module, traffic sign detection module, the detection of surface mark detection module and produces the speed control information, described rate control module is transferred to described control desk with the speed control information, and described control desk is controlled the running velocity of described motor.The direction control module receives the information of described initialization module, hunting module, detection of obstacles module, traffic lights detection module, traffic sign detection module, the detection of surface mark detection module, and generation direction control information, described direction control module is transferred to described control desk with the direction control information, and described control desk is controlled the direction of steering wheel.Mainboard Negotiation speed control module of the present invention and direction control module are controlled the running velocity of motor and the direction of steering wheel, realize miniature car should be taked after the perception to surrounding environment behaviour decision making, comprise the hunting walking, obstacle, observe traffic rules and regulations etc., improved the safety of travelling.
In a kind of preferred implementation of the present invention, mainboard adopts embedded x86 mainboard, has based on the miniature car of this mainboard that automated lane line following, site of road detect, certainly move, many cars are interactive and the ability of analysis vehicle target direction.In the present embodiment, mainboard has USB interface, and the hardware that its quantity can connect is as required adjusted, and can be limited and be not limited to 4 tunnels, and wherein, 2 road USB interfaces connect camera, and 1 road USB interface connects wireless network card; 1 road USB interface connects control desk.Multi-path digital I/O mouth is arranged on control desk, be used for connecting infrared distance sensor, electronic compass, the equipment such as motor and steering wheel.
Each module of mainboard of the present invention can realize the related algorithm of dynamic and static target environment perception under complicated road environment, comprises hunting algorithm, detection of obstacles algorithm, traffic lights detection algorithm, traffic sign detection algorithm, surface mark detection algorithm; Realize miniature car should be taked after the perception to surrounding environment behaviour decision making, comprise the hunting walking, obstacle, observe traffic rules and regulations etc.
The present invention also provides a kind of miniature vehicle control based on machine vision, and it comprises at least one miniature car, server and management end computer, and server is connected with the mainboard of miniature car, and server and mainboard carry out two-way information interaction; The management end computer is connected with server, and the information of management end computer reception server transmission is also carried out monitoring management by server to the operation of described miniature car.
In the present embodiment, miniature car is connected by TCP with server, also can miniature car terminal and server be coupled together by wireless network, server with communication to the management end computer, the running state of the miniature car of management end computer Real Time Monitoring, realize the monitoring management to miniature car.
The building to be based on the PC platform base of miniature vehicle control based on machine vision of the present invention realized, by vision sensor, on the miniature car test platform of high emulation of simulation, environment sensing and the intelligent behavior decision-making capability of miniature car tested.Miniature car is to be connected in the network of server by the information that the technological means such as radio communication will be obtained, then analyzed, draw strategy, thereby realize the control that is to miniature garage, thereby reach Che Yu road, car and car, car and urban network, realize interconnecting.On this miniature vehicle control, due to the experimental situation relative closure of miniature car, and there is not the problem of secure context substantially in miniature car, so the autonomous driving experiment can not be subject to the constraint of the non-technical elements such as laws and regulations.In addition with respect to the archeus vehicle, miniature car simple in structure, cheap, many cars test environment easily builds.And experimental site and environment are easily adjusted, and can carry out easily the experiment under multiple varying environment.The continuous progress that the research of miniature car is accompanied by day by day increasing the weight of of urban traffic blocking and intelligent transportation solution technology occurs, be also one of important embodiment of changing as management mode main, take car as object take hot spot region of municipal intelligent traffic, promoted advancing of China's intelligent transportation.
in a preferred embodiment of the present invention, has miniature car remote control administrative system in the management end computer, this miniature car remote control administrative system comprises intelligent vehicle termination management module and remote monitoring function module, wherein, the intelligent vehicle termination management module comprises system-based attribute configuration module, termination management module, status information primary attribute configuration module, status information customized management module, control command release module and collection pictures management module, miniature car by server respectively with system-based attribute configuration module, termination management module, status information primary attribute configuration module, status information customized management module, the control command release module is connected with collection pictures management module, system-based attribute configuration module is used for arranging the underlying parameter of miniature truck system, include but not limited to that specifically on miniature car, the camera position coordinate is demarcated, per second reads camera picture frame number, communication interface, the information such as miniature car network numbering.Status information primary attribute configuration module is connected with the mainboard of miniature car, is used for managing the multidate information of miniature car driving process, comprises the information such as moving velocity, acceleration-deceleration, uphill/downhill angle recognition, traveling-position coordinate.Status information customized management module is connected with miniature car owner's plate, is used for the message customization of managerial demand and remote management terminal, is used for encapsulation and also sends in real time the telemanagement parameter of miniature car.The control command release module is used for all receptions of configuration and sends form and the design parameter of instruction.Gathering the pictures management module, to be used for the management vision sensor be the road information of Real-time Collection at driving process, picture is carried out coherent analysis processes, and with result feedback to status information primary attribute configuration module.the remote monitoring function module comprises user's coupled condition monitoring module and auto-returned state or lastest imformation module, wherein, user's coupled condition monitoring module is connected with miniature car owner's plate, for realizing the be connected monitoring of every miniature vehicle control with the remote management terminal computer, the connection that shows every miniature car and management end computer, show the parameters such as speed that every miniature car is current and running state, auto-returned state or lastest imformation module are connected with miniature car owner's plate, be used for the parameters such as the miniature car running state of Real-time Obtaining, at first send reading command, miniature car sends to remote control terminal with relevant information after receiving reading command, remote control terminal upgrades data after receiving up-to-date motion state parameters, show by user's coupled condition monitoring module again.
The present invention is by the running state of the miniature car of miniature car remote control administrative system Real Time Monitoring, miniature car terminal keeps being connected with server, realization on the information network platform according to different functional requirements to miniature car attribute information and quiet, multidate information extracts and effectively utilize, and according to different functional requirements, the running state of miniature car is carried out the actv. supervision and integrated service is provided.
Patrolling method of the present invention comprises the steps:
S1: image collection module is obtained the RGB coloured image on road surface, contains lane mark information in described RGB coloured image;
S2: image processing module is converted into gray level image with described RGB coloured image;
S3: binarization block is obtained the optimal dynamic threshold value of every two field picture in described gray level image and is carried out image segmentation, obtains bianry image, and lane mark is separated;
S4: the rim detection module is carried out rim detection to described binary image, obtains containing the edge image at the inside and outside edge of lane mark;
S5: treater utilizes Hough transformation inspection vehicle diatom and obtain the lane mark parameter in described edge image, sets up the lane mark model;
S6: the described lane mark parameter that the treater utilization is obtained, obtain vehicle residing lane position data in world coordinate system by inverse perspective mapping,, according to described pavement image information, differentiate the driving mode of vehicle;
S7: vehicle corner and distance parameter that treater utilizes step S6 to obtain,, according to the driving mode of vehicle, adopt the sectional type ADAPTIVE CONTROL, send control command to controller;
S8: described controller receives the control command of treater,, by steering wheel, motor are carried out parameter adjustment, controls in real time direction of traffic and the road speed of vehicle.
In the present embodiment, the lane mark model of setting up in described step S5, in the situation that track is straight way, lane mark is straight line model; In the situation that track is bend, lane mark is the tangent line model of bend.
In the present embodiment, the driving mode in described step S6 comprises the straight way pattern, bend pattern, upward slope pattern, descending pattern, cross roads pattern, T-shaped road junction pattern, follow the mode ，Huan road pattern, nine kinds of car-parking models.
In a kind of preferred implementation of the present invention, this miniature car control method specifically comprises the steps:
The first step: build the miniature vehicle control based on machine vision of the present invention.Have this miniature car automated lane line following, site of road detect, certainly move, many cars are interactive and analyze the ability of vehicle target direction.This miniature vehicle control, based on the local area wireless network Long-distance Control, has the ability that remote vehicle tracking, inter-vehicle communication and long-range motoring condition are analyzed.In the present embodiment, according to the miniature ratio of miniature intelligent vehicle with true car 1:10, according to the height of vehicle body, the position of camera on miniature intelligent vehicle is set, make it have best blind area (headstock and the distance of obtaining the image lower edge), in the present embodiment, be taken as 28cm.
integrated use of the present invention machine vision, artificial intelligence, pattern-recognition, the pioneering technology of the multidisciplinary intersection such as wireless sense network and Meter Reliability, for the city traffic characteristics, utilize simulation technology, the real traffic circulation situation of reproduction on 3 D stereo traffic sand table simulation test platform, or fictionalize the situation of future transportation operation, make it possible to low cost, low jeopardously manifesting occurred or nonevent traffic events, its feature and rule are studied, technical scheme of the present invention is transplanted on true car, can help enterprise development to have the intelligent vehicle of complete independent intellectual property right, the driving supplementary means of controlling vehicle and prevention dangerous situation is provided for the driver, promote the vehicle control ability of navigating mate, prevention traffic accident and protection pedestrains safety.
Intelligent vehicle of the present invention is driven main research integral body automatically or as DAS (Driver Assistant System), is completed the vehicular drive task.These tasks comprise the tracking road, keep Vehicle Driving Cycle on correct road, keep a safety distance between vehicle,, according to the speed of current traffic and roadway characteristic adjusting vehicle, across track, to reach, overtake other vehicles and keep away the purpose of barrier and find the shortest path that reaches destination and travel easily in urban district and stop.Thereby the miniature intelligent vehicle based on machine vision reaches unpiloted purpose in the identification that realizes obstacle, friendship traffic lights, traffic sign etc., all will realize by machine vision.Machine vision replaces human eye to do measurement and judgement with machine exactly.
In the present embodiment, vision sensor will be ingested target and convert picture signal to, send mainboard to, and this mainboard also comprises analog-digital commutator, analog-digital commutator, according to pixel distribution and the information such as brightness, color of picture signal, is transformed into digitized signal with picture signal; The hunting module of mainboard, the detection of obstacles module, the traffic lights detection module, traffic sign detection module, surface mark detection module carry out computing to these digital signals and extract target signature separately and according to the result of differentiating, control on-the-spot device action.In the present embodiment, preferably adopt the function in the OpenCV image processing software to process image.In the present embodiment, the function that adopts includes but not limited to picture format transfer function cvCvtColor (), interesting image regions function cvSetImageROI () is set, binary conversion treatment function cvThreshold (), find profile function cvFindContours (), profile bounding box return function cvBoundingRect () etc.
Second step: the initialization module of mainboard carries out initialization to vision sensor, motor and steering wheel.
The attainable function of this module of initialization is: to the initialization of parameter in all modules, include but not limited to visual attribute sensor, projection matrix, motor speed, the direction of steering wheel is carried out initialization, in a preferred embodiment of the present invention, initialized value can be set according to experiment or this area frequently-used data.On specific algorithm is realized, mainly call the function of opencv and process.The function that the visual attribute sensor parameter call for example is set comprises:
CvCreateCameraCapture (); This function is that camera obtains function, can be obtained by this function the property value of two cameras in miniature intelligent vehicle up and down.
cvSetCaptureProperty(pCapture,CV_CAP_PROP_FRAME_WIDTH,320);
cvSetCaptureProperty(pCapture,CV_CAP_PROP_FRAME_HEIGHT,240);
These two functions are that the size of picture that camera is obtained is set, and the dimension of picture of setting in function is the size of 320 pixel * 240 pixels.
The 3rd step: vision sensor obtains the road surface picture and the road surface picture is transferred to mainboard.
The 4th the step: mainboard according to the road surface picture carry out hunting, line-hunting method comprises the steps:
S1: image collection module is obtained the RGB coloured image on road surface, contains lane mark information in described RGB coloured image, and on road, the color of lane mark is generally white usually, or is interrupted, or continuously; Pavement of road is black gray expandable, has a lot of social connections and is about 35cm, puts aside the impact of rainwater on the rainy day factor such as reflective in present embodiment;
S2: image processing module is converted into gray level image with described RGB coloured image;
S3: binarization block uses large Tianjin method obtain the optimal dynamic threshold value of every two field picture in described gray level image and carry out image segmentation, obtains bianry image, and lane mark is separated;
S4: the rim detection module uses the canny operator to carry out rim detection to described binary image, obtains containing the edge image at the inside and outside edge of lane mark;
S5: treater utilizes Hough transformation inspection vehicle diatom and obtain the lane mark parameter in described edge image, sets up the lane mark model, in the situation that track is straight way, lane mark is straight line model; In the situation that track is bend, lane mark is the tangent line model of bend;
S6: the described lane mark parameter that the treater utilization is obtained, obtain vehicle residing lane position data in world coordinate system by inverse perspective mapping, and be aided with the extraction of other features in image, differentiate the state model of vehicle current time, be straight way or bend as track of living in, whether enter cross roads or T-shaped road junction etc., so that the real-time line walking of subsequent control vehicle;
S7: vehicle corner and distance parameter that treater utilizes step S6 to obtain,, according to the driving mode of vehicle, adopt the sectional type ADAPTIVE CONTROL, send control command to controller;
S8: described controller receives the control command of treater,, by steering wheel, motor are carried out parameter adjustment, controls in real time direction of traffic and the road speed of vehicle.
Patrolling method of the present invention is stable, easy to control, and the line walking time can complete within 10 milliseconds, and the moving velocity of miniature intelligent vehicle can reach the 1-2 meter per second, and conversion is about 40-60km/h to true car.
in the present embodiment, treater also can utilize image conversion function hLines2 () to look for line, image conversion function cvHoughLines2 () has found the many lines in the image, some is wanted, some is undesired, in order to obtain lane mark, just must carry out conditional filtering, actual conditions can include but not limited to the distance to lane mark, slope threshold value, the distance of lane mark and slope threshold value can be chosen according to the distance of the lane mark on real road in concrete reality and the slope of lane mark, also can reduce in proportion or amplify and choose.
After finding lane mark,, for convenient miniature Che Huan road, can whether exist the condition in greenery patches to determine residing specifically which track according to the lane mark left and right, concrete De Huan road condition is,, if there is greenery patches on the right side, track, illustrate that track is the track on the right side, can only change trains left; If there is greenery patches in track left side, illustrate that track is the greenery patches of doing left side, can only change trains to the right, if the left and right sides does not all have greenery patches, the track of explanation in the middle of being, to the left or change trains can on right side.Line-hunting method of the present invention can find lane mark rapidly and accurately, has improved the safety of driving.
In a kind of preferred implementation of the present invention, corresponding processing is identified and made to the detection of obstacles module to the obstacle that the place ahead occurs, obstacle detection method comprises the steps:
S21: after the road surface picture that detection of obstacles module reception vision sensor obtains, the road pavement picture carries out format conversion.After the detection of obstacles module is obtained picture from vision sensor, in order to avoid the light impact, need to select suitable color space in the treating process of picture, in the present embodiment, rgb format is converted to the HSV form to be processed, the format conversion function that adopts is: cvCvtColor (image, imgHSV, CV_RGB2HSV).
S22: the detection of obstacles module is carried out gradation conversion, binary conversion treatment, searching profile and conditional filtering.In the present embodiment, call built-in function road pavement picture in opencv and carry out gradation conversion, binary conversion treatment, searching profile, conditional filtering.The concrete opencv function that mainly calls is:
Binary conversion treatment function cvThreshold ();
Find profile function cvFindContours ().
After searching out profile, obstacle is carried out conditional filtering, specifically according to obstacle, apart from features such as the color of the distance of miniature car, obstacle, size, areas, screen.Just can realize the identification to obstacle, thereby command the action of miniature car, comprise and stop the Huo Huan road.In the present embodiment, the condition of screening can be set according to actual tests.
In another kind of preferred implementation of the present invention, the traffic lights method of inspection comprises the steps:
S31: when miniature car travelled to the cross roads pattern, at first the traffic lights detection module judged whether the road picture exists stop line,, if exist, performed step S32.
S32: the height threshold according to vision sensor and traffic lights, obtain area-of-interest.In the testing process of traffic lights,, because the height of camera and traffic lights is all fixed,, in order to improve processing speed, reduce the factors such as ambient interference, therefore adopt area-of-interest is set.In the present embodiment, the height threshold of vision sensor and traffic lights can arrange according to the height of the vision sensor in concrete test or real road and traffic lights, the certain formed scope of degree of height plus-minus that the concrete height threshold that arranges is traffic lights.
S33: read R, G, the B tristimulus values of the pixel in area-of-interest and with R, G, the B tristimulus values of the traffic lights of setting, compare, when meeting error requirements, area-of-interest is target area.Concrete error limit can specifically be set according to concrete experiment.
S34: conditional filtering is carried out in target area, and the project of described conditional filtering comprises the number of screening pixel, and R, G, the shared ratio of B tristimulus values difference,, when all conditions all meets, judge red light or green light.
Traffic lights method of inspection of the present invention is by selected area-of-interest, and select target is regional in area-of-interest, has improved the rapidity that traffic lights detect, and conditional filtering is carried out in target area, has improved the accuracy that traffic lights detect.
In a kind of preferred implementation of the present invention, the traffic sign type comprises craspedodrome, forbids keeping straight on, and turns right, and no right turn, turn left, six types of " no left turn "s, and miniature intelligent vehicle is by the Discern and judge of this module realization to above traffic sign.Method for traffic sign detection comprises the steps:
S41: described traffic sign detection module detect in the picture that obtains certain pixel and on every side connected domain whether have the red pixel point, if exist, be set to target area and enter step S42, the size in concrete UNICOM territory can be set according to the size of traffic sign in test, specifically can but be not limited to less than, more than or equal to the size of traffic sign.
S42: search for black pixel point in target area, and obtain the rectangle of the connect domain that comprises black picture element, adjust the size of described rectangle by the size adjusting function, in the present embodiment, by size adjusting function cvResize (), rectangle size is adjusted into the rectangle of 5 pixel sizes of 7 pixel *.
S43: extract the characteristic information in described rectangle, with template, mate,, if the match is successful, draw the type of traffic sign, make miniature car will carry out corresponding operation.
Method for traffic sign detection of the present invention, by the size of rectangle in selected target zone and adjustment aim zone, has improved the rapidity that traffic sign detects, and characteristic information and template are mated, and has improved the accuracy that traffic sign detects.
In a kind of preferred implementation of the present invention, realize identification and the reflex action of miniature car to track of living in surface mark, as the sign of directly walking to turn right, the surface mark method of inspection comprises the steps:
S51: after the road surface picture that surface mark detection module reception vision sensor obtains, the road pavement picture carries out gradation conversion and binary conversion treatment.
S52: the surface mark detection module carries out rim detection to binary image, specifically can but be not limited to adopt Canny function binary image to carry out rim detection.
S53: the surface mark detection module looks for line to process, and according to the size of lines, screens.
S54: extract the characteristic information of described lines, with template, mate,, if the match is successful, draw the type of surface mark, make miniature car will carry out corresponding operation.
In a preferred embodiment of the present invention, the operational process of the control program of mainboard is: at first, the initialization module of mainboard carries out initialization to vision sensor, motor and steering wheel.After initialization, vision sensor detects the road surface picture and the road surface picture is transferred to hunting module, detection of obstacles module, traffic lights detection module, traffic sign detection module and surface mark detection module.The hunting module is found lane mark, and miniature car is normally advanced.In the process of advancing, the detection of obstacles module judges whether track is clean, and whether the traffic lights detection module detects uses stop line.Detection of obstacles module judgement track is clean, miniature car continuation hunting is advanced, if detection of obstacles module judgement track is unclean, the detection of obstacles module detects and passes through conditional filtering and judged whether obstacle,, if obstacle is arranged stop, again detects whether obstacle is arranged after parking, stop if still have obstacle continue, if there is no obstacle continue hunting and advance, avoided the error detection obstacle to cause the situation of stopping to occur, improved accuracy; If the judgement of detection of obstacles module does not have obstacle or does not need after conditional filtering to stop, judge whether change condition meets, if the condition of changing meets, whether the process of changing the ，Zai Huan road constantly detects to change and finishes, if finish, continue hunting and advance,, if do not finish, continue to change,, if the condition of changing does not meet, in former track, continue to travel.
The traffic lights detection module detects whether use stop line, if there is no stop line, miniature car continuation hunting is advanced, and, if stop line is arranged, has judged whether red light; If red light is arranged, stop, after red light goes out, enter the cross roads traveling mode, the hunting module is found lane mark, and after searching out lane mark, miniature car is normally advanced, if do not search out lane mark, the hunting module continues hunting; If there is no red light, the traffic sign detection module detects whether traffic sign is arranged, if having, according to the traffic sign walking, the hunting module is found lane mark, after searching out lane mark, miniature car is normally advanced, if do not search out lane mark, the hunting module continues hunting.If there is no traffic sign, enter the cross roads traveling mode, the hunting module is found lane mark, and after searching out lane mark, miniature car is normally advanced, if do not search out lane mark, the hunting module continues hunting.
The 5th step: mainboard is assigned control command to control desk, controls the operation of motor and steering wheel.Specifically the rate control module of mainboard receives the information of initialization module, hunting module, detection of obstacles module, traffic lights detection module, traffic sign detection module, the detection of surface mark detection module and produces the speed control information, described rate control module is transferred to described control desk with the speed control information, and described control desk is controlled the running velocity of described motor.Described direction control module receives the information of described initialization module, hunting module, detection of obstacles module, traffic lights detection module, traffic sign detection module, the detection of surface mark detection module, and generation direction control information, described direction control module is transferred to described control desk with the direction control information, and described control desk is controlled the direction of described steering wheel.
The 6th step: mainboard is transferred to the operation information of miniature car server and by server, is transferred to the management end computer, and the management end computer expert crosses miniature car remote control administrative system and realizes monitoring management is carried out in the operation of miniature car.
In the description of this specification sheets, the description of reference term " embodiment ", " some embodiment ", " example ", " concrete example " or " some examples " etc. means to be contained at least one embodiment of the present invention or example in conjunction with specific features, structure, material or the characteristics of this embodiment or example description.In this manual, the schematic statement of above-mentioned term not necessarily referred to identical embodiment or example.And the specific features of description, structure, material or characteristics can be with suitable mode combinations in any one or more embodiment or example.
Although illustrated and described embodiments of the invention, those having ordinary skill in the art will appreciate that: in the situation that do not break away from principle of the present invention and aim can be carried out multiple variation, modification, replacement and modification to these embodiment, scope of the present invention is limited by claim and equivalent thereof.
Claims (5)
1. the method for the vehicle obstacle-avoidance based on machine vision, is characterized in that, comprises the steps:
S1: processor adopting judges based on the detection of obstacles algorithm of single image whether miniature front side exists obstacle, if obstacle is arranged, with the detection communication in obstacle distance and obstacle track of living in to controller, and execution step S2,, if there is no obstacle, continue execution step S1;
S2: treater is demarcated two cameras, adopts the method for stereovision to judge the three-dimensional coordinate of obstacle center, center, coboundary, thus determine the height of obstacle and with described obstacle height communication to controller;
S3: the distance measuring equipment (DME) of vehicle body both sides detects the vehicle condition in adjacent two tracks, for the Bi Zhanghuan road provides the zone of wheeled and with described, exercises regional communication to described controller;
S4: controller is according to the exercised area information in obstacle distance, the obstacle track of living in, obstacle height and adjacent two tracks that obtain, adopt self adaptation to change strategy, send order to operation control module, control turning to and speed of wheel, complete independently changing of vehicle.
2. the method for the vehicle obstacle-avoidance based on machine vision as claimed in claim 1, it is characterized in that: the step of the detection of obstacles algorithm of described single image is:
S21: image is carried out colour space transformation RGB turn HSV, and extract separately the S channel image;
S22: the large Tianjin of S imagery exploitation method is carried out the dynamic threshold binaryzation obtain bianry image Bin,, to the Bin image denoising, eliminate and disturb;
S23: the obstacle profile that search institute likely exists on the Bin image, the obstacle rejecting that will judge by accident according to the screening conditions that arrange;
S24: and according to the direction deciding obstacle track of living in of lane mark.
3. the method for the vehicle obstacle-avoidance based on machine vision as claimed in claim 1, it is characterized in that: the method for described stereovision is:
S31:, according to the installation site of main camera A and auxiliary camera B, demarcate respectively, draw the intrinsic parameter of camera, outer parameter, and the spacing b between main camera A and auxiliary camera B;
S32: the image that auxiliary camera B is obtained adopts the obstacle detection method based on single image, obtains obstacle center, the two-dimensional coordinate of center, coboundary under image coordinate system;
S33: utilize following formula to obtain the three-dimensional coordinate x of obstacle under main camera system of axes
c, y
c, z
c, and utilize the calibrating parameters of camera, draw the three-dimensional coordinate of obstacle under world coordinate system,
Wherein, b is the spacing of two cameras, and f is the focal length of camera, (u
1, v
1) be the projected position of calibration point on A camera plane, (u
2, v
2) be the projected position of calibration point on B camera plane, and v
1=v
2
4. the method for the vehicle obstacle-avoidance based on machine vision as claimed in claim 1, it is characterized in that: described sectional type ADAPTIVE CONTROL refers to: deviation angle and the distance middle apart from track according to vehicle body, and vehicle body moving velocity this moment, adjust next vehicle body corner and speed of a motor vehicle constantly according to parameter d egree_per_dist self adaptation appropriateness, when near the centre of Vehicle Driving Cycle at straight way during position, remain the direction running that vehicle body is kept straight at full speed, if any side of deflection both sides, track, finely tune.
5. one kind is utilized the system of the vehicle obstacle-avoidance based on machine vision claimed in claim 1, it is characterized in that, comprises main camera A, auxiliary camera B, distance measuring equipment (DME), treater and controller,
Described main camera A and auxiliary camera B be respectively used to obtain the image on road surface and with described image transmitting to described treater;
Described treater be used for judging whether miniature front side exists obstacle and with the detection communication in obstacle distance and obstacle track of living in to controller, simultaneous processor is demarcated main camera A and auxiliary camera B, judge the three-dimensional coordinate of obstacle center, center, coboundary, thus determine the height of obstacle and with described obstacle height communication to controller;
Described distance measuring equipment (DME) is positioned at the vehicle body both sides, for detection of the vehicle condition in adjacent two tracks, also can exercise the communication in zone to described controller;
Controller is connected with operation control module, described operation control module is connected with wheel, controller is according to the exercised area information in obstacle distance, the obstacle track of living in, obstacle height and adjacent two tracks that obtain, send order to operation control module, described operation control module is controlled turning to and speed of wheel, completes independently changing of vehicle.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310333781.1A CN103386975B (en) | 2013-08-02 | 2013-08-02 | A kind of vehicle obstacle-avoidance method and system based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310333781.1A CN103386975B (en) | 2013-08-02 | 2013-08-02 | A kind of vehicle obstacle-avoidance method and system based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103386975A true CN103386975A (en) | 2013-11-13 |
| CN103386975B CN103386975B (en) | 2015-11-25 |
Family
ID=49531445
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310333781.1A Expired - Fee Related CN103386975B (en) | 2013-08-02 | 2013-08-02 | A kind of vehicle obstacle-avoidance method and system based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103386975B (en) |
Cited By (40)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104182756A (en) * | 2014-09-05 | 2014-12-03 | 大连理工大学 | Method for detecting barriers in front of vehicles on basis of monocular vision |
| CN104732835A (en) * | 2015-02-02 | 2015-06-24 | 上海交通大学 | Stadium intelligent microscopic vehicle teaching device |
| CN104833364A (en) * | 2015-05-07 | 2015-08-12 | 苏州天鸣信息科技有限公司 | Safe route indicating method for bumpy roads |
| CN105137970A (en) * | 2015-07-31 | 2015-12-09 | 奇瑞汽车股份有限公司 | Obstacle avoidance method and device for vehicle |
| CN105512628A (en) * | 2015-12-07 | 2016-04-20 | 北京航空航天大学 | Vehicle environment sensing system and method based on unmanned plane |
| CN105620409A (en) * | 2014-11-24 | 2016-06-01 | 福特全球技术公司 | Vehicle underside impact avoidance |
| CN105653257A (en) * | 2015-08-13 | 2016-06-08 | 哈尔滨安天科技股份有限公司 | Sand table system with customized strategy |
| CN105793910A (en) * | 2014-01-29 | 2016-07-20 | 爱信艾达株式会社 | Automatic driving assistance device, automatic driving assistance method, and program |
| CN105843229A (en) * | 2016-05-17 | 2016-08-10 | 中外合资沃得重工（中国）有限公司 | Unmanned intelligent vehicle and control method |
| CN106020204A (en) * | 2016-07-21 | 2016-10-12 | 触景无限科技（北京）有限公司 | Obstacle detection device, robot and obstacle avoidance system |
| CN106493748A (en) * | 2016-11-23 | 2017-03-15 | 河池学院 | A kind of robot CAS |
| CN106970616A (en) * | 2017-03-30 | 2017-07-21 | 南通大学 | A kind of intelligent tracking system |
| CN107085869A (en) * | 2017-01-16 | 2017-08-22 | 田桂昌 | Data handling system based on Internet of Things and cloud computing |
| CN107305380A (en) * | 2016-04-20 | 2017-10-31 | 上海慧流云计算科技有限公司 | A kind of automatic obstacle-avoiding method and apparatus |
| CN107618506A (en) * | 2017-09-06 | 2018-01-23 | 深圳市招科智控科技有限公司 | A kind of servomechanism obstacle avoidance system and its barrier-avoiding method |
| CN108572663A (en) * | 2017-03-08 | 2018-09-25 | 通用汽车环球科技运作有限责任公司 | Target following |
| CN108629281A (en) * | 2018-03-28 | 2018-10-09 | 深圳市路畅智能科技有限公司 | Utilize the automatic safe driving assistance method of bend corner mirror |
| CN108639065A (en) * | 2018-05-15 | 2018-10-12 | 辽宁工业大学 | A kind of vehicle safe driving control method of view-based access control model |
| CN109017786A (en) * | 2018-08-09 | 2018-12-18 | 北京智行者科技有限公司 | Vehicle obstacle-avoidance method |
| CN109214980A (en) * | 2017-07-04 | 2019-01-15 | 百度在线网络技术（北京）有限公司 | A kind of 3 d pose estimation method, device, equipment and computer storage medium |
| CN109472975A (en) * | 2017-09-08 | 2019-03-15 | 本田技研工业株式会社 | Driving assist system, drive supporting device and driving support method |
| CN109974686A (en) * | 2017-12-28 | 2019-07-05 | 沈阳新松机器人自动化股份有限公司 | Transfer robot path planning householder method based on monitoring camera detection |
| CN109987092A (en) * | 2017-12-28 | 2019-07-09 | 郑州宇通客车股份有限公司 | A kind of determination method on vehicle obstacle-avoidance lane-change opportunity and the control method of avoidance lane-change |
| WO2019144298A1 (en) * | 2018-01-23 | 2019-08-01 | 深圳市大疆创新科技有限公司 | Auxiliary movement method, mobile device and movable platform |
| CN110466512A (en) * | 2019-07-25 | 2019-11-19 | 东软睿驰汽车技术（沈阳）有限公司 | A kind of vehicle lane change method, apparatus and system |
| CN110796705A (en) * | 2019-10-23 | 2020-02-14 | 北京百度网讯科技有限公司 | Error elimination method, device, equipment and computer readable storage medium |
| CN110928301A (en) * | 2019-11-19 | 2020-03-27 | 北京小米智能科技有限公司 | Method, device and medium for detecting tiny obstacles |
| CN111497741A (en) * | 2019-01-30 | 2020-08-07 | 杭州海康威视数字技术股份有限公司 | Collision early warning method and device |
| CN111937002A (en) * | 2018-04-16 | 2020-11-13 | 三菱电机株式会社 | Obstacle detection device, automatic braking device using obstacle detection device, obstacle detection method, and automatic braking method using obstacle detection method |
| CN111951604A (en) * | 2019-04-29 | 2020-11-17 | 北京百度网讯科技有限公司 | Vehicle speed determination method, device, equipment and storage medium |
| NL2024662B1 (en) * | 2019-12-04 | 2021-04-20 | Univ Anhui Sci & Technology | Machine vision-based robot line-tracking navigation system |
| CN112819888A (en) * | 2019-11-15 | 2021-05-18 | 株式会社东芝 | Position estimation device, position estimation method, and program |
| CN112989883A (en) * | 2019-12-16 | 2021-06-18 | 中国科学院沈阳计算技术研究所有限公司 | Method for identifying obstacle in front of train |
| CN113341824A (en) * | 2021-06-17 | 2021-09-03 | 鄂尔多斯市普渡科技有限公司 | Open type automatic driving obstacle avoidance control system and control method |
| CN113574524A (en) * | 2018-10-18 | 2021-10-29 | 自动智能科技有限公司 | Method and system for obstacle detection |
| CN113788014A (en) * | 2021-10-09 | 2021-12-14 | 华东理工大学 | Special vehicle avoidance method and system based on repulsive force field model |
| CN113928217A (en) * | 2021-10-13 | 2022-01-14 | 南斗六星系统集成有限公司 | Early warning system, method and device for blocked opening of vehicle trunk and vehicle |
| CN114445335A (en) * | 2021-12-22 | 2022-05-06 | 武汉易思达科技有限公司 | Vehicle running state monitoring method and system based on binocular machine vision |
| CN115139788A (en) * | 2021-03-30 | 2022-10-04 | 本田技研工业株式会社 | Driving support system, driving support method, and storage medium |
| WO2023093306A1 (en) * | 2021-11-24 | 2023-06-01 | 上海安亭地平线智能交通技术有限公司 | Vehicle lane change control method and apparatus, electronic device, and storage medium |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110491156A (en) * | 2019-08-27 | 2019-11-22 | 无锡物联网创新中心有限公司 | A kind of cognitive method, apparatus and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006260217A (en) * | 2005-03-17 | 2006-09-28 | Advics:Kk | Traveling support device for vehicle |
| DE102011016217A1 (en) * | 2011-04-06 | 2012-10-11 | Connaught Electronics Ltd. | Method for warning driver of passenger car before rear end collision, involves projecting path of image, and warning driver of vehicle based on image with projected path, where path is determined based on velocity and steering angle |
| CN102874196A (en) * | 2011-07-11 | 2013-01-16 | 北京新岸线移动多媒体技术有限公司 | Machine vision-based automobile anti-collision method and system |
| CN103186771A (en) * | 2011-12-27 | 2013-07-03 | 哈曼（中国）投资有限公司 | Method of detecting an obstacle and driver assist system |
| CN103204163A (en) * | 2012-01-17 | 2013-07-17 | 福特全球技术公司 | Autonomous Lane Control System |
-
2013
- 2013-08-02 CN CN201310333781.1A patent/CN103386975B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006260217A (en) * | 2005-03-17 | 2006-09-28 | Advics:Kk | Traveling support device for vehicle |
| DE102011016217A1 (en) * | 2011-04-06 | 2012-10-11 | Connaught Electronics Ltd. | Method for warning driver of passenger car before rear end collision, involves projecting path of image, and warning driver of vehicle based on image with projected path, where path is determined based on velocity and steering angle |
| CN102874196A (en) * | 2011-07-11 | 2013-01-16 | 北京新岸线移动多媒体技术有限公司 | Machine vision-based automobile anti-collision method and system |
| CN103186771A (en) * | 2011-12-27 | 2013-07-03 | 哈曼（中国）投资有限公司 | Method of detecting an obstacle and driver assist system |
| CN103204163A (en) * | 2012-01-17 | 2013-07-17 | 福特全球技术公司 | Autonomous Lane Control System |
Cited By (55)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105793910A (en) * | 2014-01-29 | 2016-07-20 | 爱信艾达株式会社 | Automatic driving assistance device, automatic driving assistance method, and program |
| US9836051B2 (en) | 2014-01-29 | 2017-12-05 | Aisin Aw Co., Ltd. | Automated drive assisting device, automated drive assisting method, and program |
| CN105793910B (en) * | 2014-01-29 | 2018-01-19 | 爱信艾达株式会社 | Automatic Pilot servicing unit, automatic Pilot householder method |
| CN104182756B (en) * | 2014-09-05 | 2017-04-12 | 大连理工大学 | Method for detecting barriers in front of vehicles on basis of monocular vision |
| CN104182756A (en) * | 2014-09-05 | 2014-12-03 | 大连理工大学 | Method for detecting barriers in front of vehicles on basis of monocular vision |
| CN105620409A (en) * | 2014-11-24 | 2016-06-01 | 福特全球技术公司 | Vehicle underside impact avoidance |
| US10183659B2 (en) | 2014-11-24 | 2019-01-22 | Ford Global Technologies, Llc | Vehicle underside impact avoidance |
| CN104732835A (en) * | 2015-02-02 | 2015-06-24 | 上海交通大学 | Stadium intelligent microscopic vehicle teaching device |
| CN104732835B (en) * | 2015-02-02 | 2017-08-01 | 上海交通大学 | A kind of intelligent miniature vehicle instructional device in stadium |
| CN104833364A (en) * | 2015-05-07 | 2015-08-12 | 苏州天鸣信息科技有限公司 | Safe route indicating method for bumpy roads |
| CN104833364B (en) * | 2015-05-07 | 2018-05-18 | 深圳市爱民科技有限公司 | A kind of emergency route indicating means on bump course |
| CN105137970A (en) * | 2015-07-31 | 2015-12-09 | 奇瑞汽车股份有限公司 | Obstacle avoidance method and device for vehicle |
| CN105137970B (en) * | 2015-07-31 | 2018-03-16 | 奇瑞汽车股份有限公司 | Vehicle obstacle-avoidance method and device |
| CN105653257B (en) * | 2015-08-13 | 2023-04-07 | 安天科技集团股份有限公司 | Sand table system capable of customizing strategy |
| CN105653257A (en) * | 2015-08-13 | 2016-06-08 | 哈尔滨安天科技股份有限公司 | Sand table system with customized strategy |
| CN105512628B (en) * | 2015-12-07 | 2018-10-23 | 北京航空航天大学 | Vehicle environmental sensory perceptual system based on unmanned plane and method |
| CN105512628A (en) * | 2015-12-07 | 2016-04-20 | 北京航空航天大学 | Vehicle environment sensing system and method based on unmanned plane |
| CN107305380A (en) * | 2016-04-20 | 2017-10-31 | 上海慧流云计算科技有限公司 | A kind of automatic obstacle-avoiding method and apparatus |
| CN105843229A (en) * | 2016-05-17 | 2016-08-10 | 中外合资沃得重工（中国）有限公司 | Unmanned intelligent vehicle and control method |
| CN106020204A (en) * | 2016-07-21 | 2016-10-12 | 触景无限科技（北京）有限公司 | Obstacle detection device, robot and obstacle avoidance system |
| CN106493748A (en) * | 2016-11-23 | 2017-03-15 | 河池学院 | A kind of robot CAS |
| CN107085869A (en) * | 2017-01-16 | 2017-08-22 | 田桂昌 | Data handling system based on Internet of Things and cloud computing |
| CN108572663A (en) * | 2017-03-08 | 2018-09-25 | 通用汽车环球科技运作有限责任公司 | Target following |
| CN106970616A (en) * | 2017-03-30 | 2017-07-21 | 南通大学 | A kind of intelligent tracking system |
| CN109214980A (en) * | 2017-07-04 | 2019-01-15 | 百度在线网络技术（北京）有限公司 | A kind of 3 d pose estimation method, device, equipment and computer storage medium |
| CN107618506A (en) * | 2017-09-06 | 2018-01-23 | 深圳市招科智控科技有限公司 | A kind of servomechanism obstacle avoidance system and its barrier-avoiding method |
| CN107618506B (en) * | 2017-09-06 | 2021-02-23 | 深圳市招科智控科技有限公司 | Obstacle avoidance system for automatic driving device and obstacle avoidance method thereof |
| CN109472975A (en) * | 2017-09-08 | 2019-03-15 | 本田技研工业株式会社 | Driving assist system, drive supporting device and driving support method |
| CN109987092B (en) * | 2017-12-28 | 2020-10-30 | 郑州宇通客车股份有限公司 | Method for determining vehicle obstacle avoidance and lane change time and method for controlling obstacle avoidance and lane change |
| CN109974686A (en) * | 2017-12-28 | 2019-07-05 | 沈阳新松机器人自动化股份有限公司 | Transfer robot path planning householder method based on monitoring camera detection |
| CN109987092A (en) * | 2017-12-28 | 2019-07-09 | 郑州宇通客车股份有限公司 | A kind of determination method on vehicle obstacle-avoidance lane-change opportunity and the control method of avoidance lane-change |
| WO2019144298A1 (en) * | 2018-01-23 | 2019-08-01 | 深圳市大疆创新科技有限公司 | Auxiliary movement method, mobile device and movable platform |
| CN108629281A (en) * | 2018-03-28 | 2018-10-09 | 深圳市路畅智能科技有限公司 | Utilize the automatic safe driving assistance method of bend corner mirror |
| CN111937002A (en) * | 2018-04-16 | 2020-11-13 | 三菱电机株式会社 | Obstacle detection device, automatic braking device using obstacle detection device, obstacle detection method, and automatic braking method using obstacle detection method |
| CN108639065A (en) * | 2018-05-15 | 2018-10-12 | 辽宁工业大学 | A kind of vehicle safe driving control method of view-based access control model |
| CN109017786A (en) * | 2018-08-09 | 2018-12-18 | 北京智行者科技有限公司 | Vehicle obstacle-avoidance method |
| CN113574524A (en) * | 2018-10-18 | 2021-10-29 | 自动智能科技有限公司 | Method and system for obstacle detection |
| CN111497741A (en) * | 2019-01-30 | 2020-08-07 | 杭州海康威视数字技术股份有限公司 | Collision early warning method and device |
| CN111951604B (en) * | 2019-04-29 | 2022-05-10 | 北京百度网讯科技有限公司 | Vehicle speed determination method, device, equipment and storage medium |
| CN111951604A (en) * | 2019-04-29 | 2020-11-17 | 北京百度网讯科技有限公司 | Vehicle speed determination method, device, equipment and storage medium |
| CN110466512A (en) * | 2019-07-25 | 2019-11-19 | 东软睿驰汽车技术（沈阳）有限公司 | A kind of vehicle lane change method, apparatus and system |
| CN110796705A (en) * | 2019-10-23 | 2020-02-14 | 北京百度网讯科技有限公司 | Error elimination method, device, equipment and computer readable storage medium |
| CN112819888A (en) * | 2019-11-15 | 2021-05-18 | 株式会社东芝 | Position estimation device, position estimation method, and program |
| CN110928301A (en) * | 2019-11-19 | 2020-03-27 | 北京小米智能科技有限公司 | Method, device and medium for detecting tiny obstacles |
| NL2024662B1 (en) * | 2019-12-04 | 2021-04-20 | Univ Anhui Sci & Technology | Machine vision-based robot line-tracking navigation system |
| CN112989883A (en) * | 2019-12-16 | 2021-06-18 | 中国科学院沈阳计算技术研究所有限公司 | Method for identifying obstacle in front of train |
| CN112989883B (en) * | 2019-12-16 | 2024-02-02 | 中国科学院沈阳计算技术研究所有限公司 | Method for identifying obstacle in front of train |
| CN115139788A (en) * | 2021-03-30 | 2022-10-04 | 本田技研工业株式会社 | Driving support system, driving support method, and storage medium |
| CN113341824A (en) * | 2021-06-17 | 2021-09-03 | 鄂尔多斯市普渡科技有限公司 | Open type automatic driving obstacle avoidance control system and control method |
| CN113788014B (en) * | 2021-10-09 | 2023-01-24 | 华东理工大学 | Special vehicle avoidance method and system based on repulsive force field model |
| CN113788014A (en) * | 2021-10-09 | 2021-12-14 | 华东理工大学 | Special vehicle avoidance method and system based on repulsive force field model |
| CN113928217A (en) * | 2021-10-13 | 2022-01-14 | 南斗六星系统集成有限公司 | Early warning system, method and device for blocked opening of vehicle trunk and vehicle |
| WO2023093306A1 (en) * | 2021-11-24 | 2023-06-01 | 上海安亭地平线智能交通技术有限公司 | Vehicle lane change control method and apparatus, electronic device, and storage medium |
| CN114445335A (en) * | 2021-12-22 | 2022-05-06 | 武汉易思达科技有限公司 | Vehicle running state monitoring method and system based on binocular machine vision |
| CN114445335B (en) * | 2021-12-22 | 2024-04-12 | 武汉易思达科技有限公司 | Vehicle running state monitoring method based on binocular machine vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103386975B (en) | 2015-11-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103386975B (en) | A kind of vehicle obstacle-avoidance method and system based on machine vision | |
| CN103389733A (en) | Vehicle line walking method and system based on machine vision | |
| CN103268072A (en) | Miniature vehicle, miniature vehicle control system and control method based on machine vision | |
| US11334753B2 (en) | Traffic signal state classification for autonomous vehicles | |
| CN108982117B (en) | Unmanned vehicle testing system and method | |
| CN108959321B (en) | Parking lot map construction method, system, mobile terminal and storage medium | |
| CN111874006B (en) | Route planning processing method and device | |
| CN109849922B (en) | Visual information and GIS information fusion-based method for intelligent vehicle | |
| CN113378947A (en) | Vehicle road cloud fusion sensing system and method for unmanned transportation in open-pit mining area | |
| DE112020004133T5 (en) | SYSTEMS AND PROCEDURES FOR IDENTIFICATION OF POSSIBLE COMMUNICATION BARRIERS | |
| CN103413313B (en) | The binocular vision navigation system of electrically-based robot and method | |
| CN108068817A (en) | A kind of automatic lane change device and method of pilotless automobile | |
| CN105512628A (en) | Vehicle environment sensing system and method based on unmanned plane | |
| CN103400392A (en) | Binocular vision navigation system and method based on inspection robot in transformer substation | |
| CN101944176A (en) | Exist the more excellent clear path of means of transportation sign to detect | |
| CN110379168A (en) | A kind of vehicular traffic information acquisition method based on Mask R-CNN | |
| CN113071518B (en) | Automatic unmanned driving method, minibus, electronic equipment and storage medium | |
| CN101702263A (en) | Pedestrian crosswalk signal lamp green wave self-adaption control system and method | |
| KR101510745B1 (en) | Autonomous vehicle system | |
| US11673581B2 (en) | Puddle occupancy grid for autonomous vehicles | |
| CN115565262A (en) | Parking charging automatic inspection robot and method | |
| CN104407612A (en) | Miniature intelligent vehicle and test method applying miniature intelligent vehicle | |
| CN105590087A (en) | Road recognition method and device | |
| CN116597690B (en) | Highway test scene generation method, equipment and medium for intelligent network-connected automobile | |
| CN115840404B (en) | Cloud control automatic driving system based on automatic driving special road network and digital twin map |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20151125 Termination date: 20180802 |