CN105654732A - Road monitoring system and method based on depth image - Google Patents
Road monitoring system and method based on depth image Download PDFInfo
- Publication number
- CN105654732A CN105654732A CN201610120682.9A CN201610120682A CN105654732A CN 105654732 A CN105654732 A CN 105654732A CN 201610120682 A CN201610120682 A CN 201610120682A CN 105654732 A CN105654732 A CN 105654732A
- Authority
- CN
- China
- Prior art keywords
- road
- unit
- monitoring
- module
- depth image
- 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
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
- G08G1/0175—Detecting movement of traffic to be counted or controlled identifying vehicles by photographing vehicles, e.g. when violating traffic rules
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of computer vision, in particular to a road monitoring system and method based on a depth image. An image collection module comprises a plurality of collection units. The multiple collection units synchronously collect road information so as to obtain a plurality of road images. A depth calculation module calculates three-dimensional information of each pixel point in the multiple road images. A recognition module recognizes a monitored target on a road to obtain recognition information. A judgment module judges recognition information according to judgment rules to obtain a judgment result. The road monitoring method comprises the steps of obtaining the multiple road images; matching homonymy points in the multiple road images, and calculating the three-dimensional information of each pixel point in the multiple road images; recognizing the monitored target on the road according to the multiple road images and the three-dimensional information to obtain recognition information; utilizing the judgment rules for judging the recognition information to obtain the judgment result, and determining whether a control instruction is output or not according to the judgment result.
Description
Technical field
The present invention relates to computer vision field, relate in particular to a kind of road based on depth imageMonitoring system and method.
Background technology
The preventing road monitoring system of extensive use at present can only simply record and store video data, nothingMethod is accomplished the warning in advance of event, obtains further informational needs and manually video is carried out to interpretation.Perfect along with the extension of road and road monitoring network, artificial interpretation more and more can not meetThe demand of urban construction and management, especially wisdom city and internet+background under, to roadRoad monitoring has proposed more and more higher requirement: not only will carry out real time record to road monitoring videoAnd storage, also need to use computer vision technique to carry out real-time analysis, generation to video contentFor manually making differentiation. For example, know through the pedestrian on road surface and the behavioural analysis of vehicle, car plateNot, vehicle speed measurement, traffic violations record, throwing object in high sky and ground foreign matter early warning etc. automatically. WithTime, the monitoring system based on single camera cannot obtain yardstick and three-dimensional information, can not be fineThe object that realizes Intelligent road monitoring.
Summary of the invention
The problem existing for prior art, the invention provides a kind of real-time, high-precision baseIn preventing road monitoring system and the method for depth image.
The present invention adopts following technical scheme:
Based on a preventing road monitoring system for depth image, described preventing road monitoring system comprises:
Image capture module, comprises multiple collecting units, described multiple collecting unit synchronous acquisitionsRoad information is to obtain multiple road images;
Depth calculation module, is connected with described image capture module, to described multiple road imagesIn same place mate, and calculate three of each pixel in described multiple road imageDimension information;
Identification module, is connected with described image capture module, described depth calculation module respectively,The monitoring objective on road is carried out according to described multiple road images and described three-dimensional informationIdentify and then obtain identifying information;
Judge module, is connected with described identification module, in described judge module, is pre-stored with judgementRule, and described judge module judges described identifying information according to described judgment ruleAnd export judged result.
Preferably, described preventing road monitoring system comprises described far-end server and terminal device, itsIn,
Described terminal device comprises described image capture module;
Described far-end server comprise described depth calculation module, described identification module and described in sentenceDisconnected module.
Preferably, described image capture module comprises multiple RGB cameras, described multiple RGBCamera is taken respectively road image, to obtain RGB image.
Preferably, described multiple collecting unit aligns and establishes in the horizontal direction of its projection centrePut.
Preferably, described depth calculation module comprises:
Acquiring unit, is connected with described image capture module, obtains described image capture module and adoptsDescribed multiple road images of collection;
Demarcate unit, be connected with described acquiring unit, adopt computer vision technique to described manyIndividual collecting unit carries out respectively the demarcation of position, obtains multiple collecting unit coordinates;
Matching unit, is connected with described acquiring unit, of the same name in described multiple road imagesPoint mates, and extracts multiple characteristic points of described monitoring objective;
Computing unit, is connected with described demarcation unit, described matching unit respectively, according to describedMultiple collecting unit coordinate utilizations recovery formula calculate respectively the three-dimensional letter of described multiple characteristic pointsBreath, to obtain the three-dimensional information of described monitoring objective.
Preferably, described matching unit is by carrying out rim detection and feature to described monitoring objectiveThe method of coupling is extracted multiple characteristic points of described monitoring objective.
Preferably, described image capture module comprises the first collecting unit and the second collecting unit,The collecting unit coordinate P of described the first collecting unit is demarcated in described demarcation unitaFor x1，y1, instituteState demarcation unit and demarcate the collecting unit coordinate P of described the second collecting unitbFor x2，y2, a spyThe three-dimensional information of levying is a little XC，YC，ZC, and the recovery formula of this characteristic point is:
Wherein, | Pa-Pb| be the road of described the first collecting unit and the collection of the second collecting unitThe position skew of image, the focal length that f is collecting unit, BCFor the throwing of described the first collecting unitThe line distance of the projection centre of shadow center and described the second collecting unit.
Preferably, described identification module comprises:
Pretreatment unit, detects monitoring objective in described road image and described monitoring objectiveThree-dimensional information;
Recognition unit, is connected with described detecting unit, by described monitoring objective and described three-dimensional letterBreath mates, identifies with image template, obtains described identifying information.
Preferably, described identification module also comprises:
Initialization unit, is connected with described recognition unit, is pre-stored with described image template.
Preferably, described identifying information comprises: the position of described monitoring objective and/or speed and/Or three-dimensional geometry yardstick.
Preferably, described preventing road monitoring system also comprises:
Memory module, respectively with described image capture module, described depth calculation module connects,Store described road image and described three-dimensional information.
Preferably, described judge module comprises:
Rule unit, is pre-stored with judgment rule;
Judging unit, judges described identifying information, obtains described judged result;
Whether output unit, is connected with described judging unit, determine defeated according to described judgment ruleGo out the control instruction of described judged result.
Preferably, described preventing road monitoring system also comprises:
Executive Module, is connected with described judge module, carries out described control instruction.
Preferably, described Executive Module comprises:
Signal transmitter unit, is connected with described judge module, sends described control instruction;
Signal receiving unit, is connected with described signal transmitter unit, receives described control instruction;
Alarm unit, is connected with described signal receiving unit, according to the described control instruction receivingReport to the police;
Remote monitoring unit, is connected with described signal receiving unit, utilizes described control instruction to enterRow is further differentiated operation.
Preferably, described alarm unit comprises:
Fixed terminal alarm unit, reports to the police in the position presetting;
Mobile terminal alarm unit, the mode of the movement of employing is reported to the police.
Based on a road monitoring method for depth image, described road monitoring method comprises:
Step S1, multiple collecting unit synchronous acquisition road informations, obtain multiple road images;
Step S2, mates the same place in described multiple road images, and calculatesThe three-dimensional information of each pixel in described multiple road image;
Step S3, according to described multiple road images and described three-dimensional information to the prison on roadControl target is identified, and obtains identifying information;
Step S4, utilizes judgment rule to judge described identifying information, obtains judgement knotReally, and determine whether according to described judged result output control instruction.
Preferably, described road monitoring method also comprises:
Step S11, before described step S1, gathers standard road image, synthetic imageTemplate.
Preferably, described step S2 specifically comprises:
Step S21, obtains described multiple road images that described image capture module gathers;
Step S22, adopts computer vision technique to carry out respectively position to described multiple collecting unitsThe demarcation of putting, obtains multiple collecting unit coordinates;
Step S23, mates the same place in described multiple road images, described in extractionMultiple characteristic points of monitoring objective;
Step S24, recovers formula according to described multiple collecting unit coordinate utilizations and calculates respectively instituteState the three-dimensional information of multiple characteristic points, to obtain the three-dimensional information of described monitoring objective.
Preferably, described step S3 specifically comprises:
Step S31, detects three of monitoring objective in described road image and described monitoring objectiveDimension information;
Step S32, by described monitoring objective mate with described image template with described three-dimensional information,Identification, obtains described identifying information.
Preferably, step S4 specifically comprises:
Step S41, pre-stored judgment rule;
Step S42, judges described identifying information, obtains described judged result;
Step S43, determines the control of whether exporting described judged result according to described judgment ruleInstruction.
Preferably, described road monitoring method also comprises:
Step S5, after described step S4, carries out described control instruction.
Preferably, described step S5 specifically comprises:
Step S51, sends described control instruction;
Step S52, receives described control instruction;
Step S53, reports to the police according to the described control instruction receiving;
Step S54, utilizes described control instruction to carry out further differentiating operation.
The invention has the beneficial effects as follows:
(1) comprehensively judge the two-dimensional signal and three on road by RGB image and depth imageDimension information, improves accuracy and the reliability of monitoring.
(2), based on computer vision technique, road is carried out to target detection, based on engineeringThe method of practising, carries out behavioural analysis to the monitoring objective in road image, and according to identifying informationBehavior is carried out to classification, the behavior that exceedes certain rank is carried out to early warning.
(3) use computer vision technique to carry out real-time analysis, generation to the content of road imageFor manual detection, judgement, avoid the erroneous judgement producing because of human factor.
Brief description of the drawings
Fig. 1 is the structural representation of a kind of preventing road monitoring system based on depth image of the present invention;
Fig. 2 is the structural representation of depth calculation module of the present invention;
Fig. 3 is the schematic diagram of binocular stereo vision three-dimensional reconstruction of the present invention;
Fig. 4 is the monitoring system figure that the present invention is based on memory module;
Fig. 5 is the monitoring system structure chart that the present invention is based on Executive Module;
Fig. 6 is the structural representation of Executive Module of the present invention;
Fig. 7 is the flow chart of a kind of road monitoring method based on depth image of the present invention.
Detailed description of the invention
It should be noted that, in the situation that not conflicting, following technical proposals, technical characterictic itBetween can mutually combine.
Below in conjunction with accompanying drawing, the specific embodiment of the present invention is further described:
The invention provides one utilizes terminal device to carry out reality to road by far-end serverTime, high-precision monitoring system and method, this system and method is in existing technical schemeOn basis, increase obtaining of three-dimensional information to monitoring objective, can be from the angle of 3-D viewMore accurate monitoring in real time, has filled up the blind area of three-dimensional observation in prior art.
Embodiment mono-
The present embodiment provides a kind of preventing road monitoring system based on depth image, as Fig. 1 instituteShow, the preventing road monitoring system of the present embodiment is intended to explanation and utilizes terminal device by far-end serverRoad is monitored, and wherein terminal device comprises image capture module, this image capture moduleComprise multiple collecting units, collecting unit can be RGB (RedGreenBlue) camera,RGB camera can be taken RGB image as road image, and wherein, RGB image adoptsRgb color pattern, a kind of color standard that RGB is industrial quarters, be by red (R),Each other folded of the variation of the three primary colors passage of green (G), blue (B) three light and theyCalais obtains color miscellaneous, and RGB represents three passages of red, green, blueColor, this standard almost comprised mankind's eyesights can perception all colours, be current fortuneWith one of the widest color system. The application scenarios of RGB image is also comparatively extensive, for example orderFront display is mostly to have adopted RGB color standard, on display, is to pass through electronicsRifle is beaten and is produced color at the red, green, blue three-colour light-emitting of screen on extremely, and computer generally also canShow 32 colors, have 10,000,000 kinds of above colors.
As mentioned above, the present embodiment describes for example taking collecting unit as RGB camera, thisIn embodiment, the number of collecting unit is exemplified as two, is convenient to so follow-up describing, i.e. figureComprise two RGB cameras as acquisition module, these two RGB cameras are gathering road image (thisIn embodiment, be illustrated the one that RGB image is road image with RGB image)After, road image need to be transferred to far-end server and process, this transmission network is eachThe internet at RGB camera and far-end server place forms a distributed regional networkPlatform, network monitoring that like this can feasible region property, also can realize global strange land prisonControl, global monitoring due to by this over distance transmission of current internet realize flatPlatform, increases the scope of the road monitoring of the present embodiment, and monitoring brings to system at a distanceBe management, facility in controls, and on image transmitting fast.
Further, as shown in Figure 1, the far-end server in the present embodiment comprises depth calculationModule, identification module, judge module, depth calculation module is connected with image capture module, usesIn passing through the road image of binocular camera (two RGB cameras) synchronous acquisition, and by roadSame place in the image of road mates, thus utilize similar triangles theorem calculate two dimensionThe three-dimensional information of each pixel in road image, and then obtain the three-dimensional information of monitoring objective,One of main inventive point that depth calculation module is the present embodiment simultaneously, below will be for depth gaugeThe course of work of calculating unit in module is described in detail.
As shown in Figure 2, depth calculation module mainly comprises following several subelement, acquiring unit,Demarcate unit, matching unit, computing unit, wherein, acquiring unit and image capture module connectConnect, acquiring unit is for obtaining RGB image from image capture module, by demarcating unit to oftenThe position of individual collecting unit is demarcated, to obtain the collecting unit coordinate of each collecting unit,The inside and outside parameter of RGB camera is just asked for by camera calibration.
Preferably, the projection centre of each collecting unit in the present embodiment is right in the horizontal directionTogether, make the baseline long enough of binocular camera, change to depth image because of Camera extrinsic thereby reduceThe error that calculating brings, as shown in Figure 3, for example binocular camera (two RGB cameras)Coordinate is respectively Pa(x1,y1)、Pb(x2,y2); Matching unit is connected with acquiring unit, and coupling is singleUnit utilizes the internal reference of binocular camera platform and outer ginseng to carry out distortion correction to the RGB image gatheringWith polar curve proofread and correct, on same polar curve use characteristic point mate find world coordinate system in same thingThe different pixels point of body on two width RGB images, to the feature of monitoring objective in RGB imagePoint extracts, and the process of extract minutiae, just completes by rim detection and characteristic matching,Obtain thus the coordinate information of characteristic point.
Computing unit respectively with demarcate unit, matching unit and be connected, for will demarcate unit,Joining unit obtains characteristic point coordinate information and collecting unit coordinate information and calculates by recovering formulaGo out the three-dimensional coordinate information (three-dimensional information) based on collecting unit coordinate system, be finally converted to byThe world coordinate system coordinate of setting voluntarily according to demand, obtains accurately complete observed object image spyLevy after real space coordinate a little, carry out depth recovery by range of triangle principle.
As shown in Figure 3, Fig. 3 is the schematic diagram of binocular stereo vision in the present embodiment, image aWith image b be the road image of same monitoring objective, performance parameters is identical, position is solid by twoFixed RGB camera is taken and is obtained from different perspectives, so its visual effect, position etc. are to some extentDifference. BCBeing baseline, is also the line distance of RGB camera projection centre, RGB camera mirrorCenter is lens focus f to the distance of camera imaging plane, RGB camera optical axis and mileage chartThe intersection point O of picture plane1、O2For plane coordinates initial point, build taking left RGB image center as initial pointVertical three-dimensional system of coordinate XC-YC-ZC。
Point P is the characteristic point on monitoring objective in space, in left RGB camera coordinates isThree-dimensional coordinate is (XC,YC,ZC), the image space on two width road images is respectively Pa(x1,y1)、Pb(x2,y2), can obtain by the geometrical relationship of similar triangles:
In formula, x1、y1、x2、y2The physical coordinates in surface road image, BCRGBThe external parameter of camera, f is the inner parameter of RGB camera, | Pa-Pb| be parallax, put PPosition skew in two secondary road images. As previously mentioned, the imaging of two RGB cameras is flatFace row after polar curve is proofreaied and correct is aimed at, imaging plane with optical axis vertical direction in same flatEvery a line of face and image is strictly alignd, therefore y=y1＝y2, parallax is | x1-x2|. Therefore,The space coordinates of 1 P in space on monitoring objective can be by parallax in conjunction with inside and outside RGB cameraCalculation of parameter draws. And the three-dimensional reconstruction process of monitoring objective is exactly the road at monitoring objectiveOn image, find out the characteristic point that can characterize in a large number monitoring objective information, calculate these spaces of points and sitTarget process.
The preventing road monitoring system of the present embodiment also comprises identification module, identification module respectively with imageAcquisition module is connected with depth calculation module, and identification module is mainly used according to three-dimensional information and roadRoad image adopts computer vision technique to identify monitoring objective, the pre-place in identification moduleReason unit adopts frame difference method and background subtraction method, extracts foreign matter and moving target in visual field,For monitoring objective, the recognition unit in identification module adopts image template coupling and the degree of depth to learnMethod is identified target and foreign matter, grades according to result for follow-up. Wherein,Identification module also comprises initialization unit, is connected with described recognition unit, above-mentioned for storingImage template, the follow-up detailed description in detail of acquisition process of the image template in initialization unit. IdentificationThe operation principle of unit is: when there is no object (monitoring objective) in overlapping region, binocular camera visual fieldOccur or through out-of-date, road image does not change, road image can not produce deformation; When lookingWhile there is foreign matter and moving target (monitoring objective) in, road image produces deformation in part,At this moment, according to RGB camera coordinates system, monitoring objective to parameters such as RGB camera plane distances,Can measure the identifying informations such as position, speed and the three-dimensional geometry yardstick of monitoring objective.
Further, the preventing road monitoring system of the present embodiment can also comprise a memory module, asShown in Fig. 4, memory module is connected with image capture module, depth calculation module respectively, storageModule is used for storing the three-dimensional information of road image and road image monitoring objective, so thatFollow-uply search, transfer.
The judge module of the present embodiment comprises regular unit, judging unit, output unit, on ruleIn unit, be pre-stored with judgment rule, judging unit is connected with regular unit, judging unit pairIdentifying information judges, obtains judged result, default the sentencing of output unit implementation rule unitDisconnected rule, and according to above-mentioned judged result, in conjunction with the rule of setting in advance, selectively willThe target monitoring and event classification are uploaded to remote monitoring unit, do further to differentiate. FarFollow-up being elaborated of the course of work of range monitoring unit.
The preventing road monitoring system of the present embodiment also comprises Executive Module, as shown in Figure 5, judges mouldOutput unit in piece utilizes the comprehensive judged result of judgment rule comprehensively to judge, and can be to respectivelyExecutive Module in control point is assigned specific control instruction, and Executive Module is held according to control instructionRow judged result. For example: on-the-spot acoustic control warning, portable set are reported to the police, Surveillance center reports to the policeDeng, in the time being necessary, can transfer road image and the three-dimensional information of each control point.
As shown in Figure 6, Executive Module comprises alarm unit and remote monitoring unit, remote monitoringUnit receives the judged result from RGB image, the high-resolution of comprehensive utilization RGB imageThe parameters such as the three-dimensional information comprising with depth image, melt the information of two parts image processingClose, can further differentiate target and event.
Alarm unit can comprise the warning of mobile terminal and the warning of fixed terminal, mobile terminalCan be the warning of the mobile devices such as mobile phone, user can learn alarm signal in position arbitrarilyBreath, the warning of fixed terminal is just in time contrary, and user need to can obtain in specific positionWarning message.
In sum, embodiment mono-provides a kind of preventing road monitoring system based on depth image,By the depth information of monitoring objective in visual field, in scene, each point is flat with respect to RGB cameraThe distance of face, thus calculate position and the three-dimensional geometry size etc. of target, and process in real timeThe problem existing on road.
Embodiment bis-
As shown in Figure 7, the present embodiment provides a kind of road monitoring side based on depth imageMethod, the preventing road monitoring system that this road monitoring method can be based in above-described embodiment one, this roadRoad method for supervising comprises:
Step S1, multiple collecting unit synchronous acquisition road informations, obtain multiple road images;
Step S2, mates the same place in multiple road images, and calculates multipleThe three-dimensional information of each pixel in road image;
Step S3, enters the monitoring objective on road according to multiple road images and three-dimensional informationRow identification, obtains identifying information;
Step S4, utilizes judgment rule to judge identifying information, obtains judged result,And determine whether export control instruction according to judged result;
Step S5, carries out control instruction.
In the present embodiment, be mainly divided into five key steps, existing respectively for five steps is described in detail in detailDescribe in detail, before step S1, also need memory image template, the source of image templateCan be to gather by image capture module, be stored in afterwards in identification module, for rearContinuous contrast.
Concrete, in step S1, image capture module obtains RGB image, RGB cameraStart working and record the road image in guarded region, then transmitting it to depth calculation mouldPiece is processed.
In step S2, depth calculation resume module RGB image, through binocular correction, binocularThe processes such as coupling utilize similar triangles theorem to obtain the three-dimensional information of RGB image, and one storesModule stores three-dimensional information and RGB image, export afterwards three-dimensional information and be uploaded to identification mouldPiece.
Wherein, depth calculation module comprises acquiring unit, demarcates unit, matching unit and calculatingUnit, acquiring unit is obtained multiple road images that image capture module gathers; Demarcation unit is adoptedWith computer vision technique, multiple collecting units are carried out respectively the demarcation of position, obtain multiple adoptingCollection unit coordinate; Matching unit mates the same place in multiple road images, extracts prisonMultiple characteristic points of control target; Computing unit recovers formula according to multiple collecting unit coordinate utilizationsCalculate respectively the three-dimensional information of multiple characteristic points, to obtain the three-dimensional information of monitoring objective, whereinComputing unit utilizes similar triangles theorem to obtain true monitoring objective to camera (RGB camera)The range information of plane. At this moment, arrive camera plane distance according to camera coordinates system, monitoring objectiveDeng parameter, can measure the identifying informations such as position, speed and the three-dimensional geometry yardstick of monitoring objective,Its computational process is identical with the process in embodiment mono-, does not repeat herein.
In step S3, identification module is processed RGB image and three-dimensional information, adopt image cut apart,The methods such as frame difference method and background subtraction method, extract foreign matter and moving target in visual field, and adoptBy the method for template matches and degree of depth study, target and foreign matter are identified, moving target is enteredRow behavioural analysis, then grades to target and behavior thereof according to identification and analysis result,To identifying information.
The specific implementation unit of step S3 is: pretreatment unit detects the monitoring in road imageThe three-dimensional information of target and monitoring objective; Recognition unit is by monitoring objective and three-dimensional information and imageTemplate matches, identification, obtain identifying information. Wherein, initialization unit, connects with recognition unitConnect, be pre-stored with image template.
In step S4, judge module, according to identifying information, is assigned execution instruction, output unitAccording to above-mentioned result, in conjunction with the rule of setting in advance, selectively by the order monitoringMark and event classification are uploaded to the server of Surveillance center, do further to differentiate.
Wherein, regular unit is pre-stored with judgment rule; Judging unit is sentenced identifying informationDisconnected, obtain judged result; Output unit determines whether to export judged result according to judgment ruleControl instruction.
In step S5, Executive Module is according to control instruction, executable operations, and Executive Module is comprehensiveThe road image of each control point (collecting unit) comprehensively judges, and can be under each control pointReach specific control instruction, for example: during on-the-spot acoustic control warning, portable set are reported to the police, monitoredHeart warnings etc. can be transferred identifying information and the road image of each control point in the time being necessary.
Wherein, signal transmitter unit sending controling instruction; Signal receiving unit receives control instruction;Alarm unit is reported to the police according to the control instruction receiving; Remote monitoring unit by using control instructionCarry out further differentiating operation.
To sum up, identification module can carry out believing based on three-dimensional to road based on computer vision techniqueBreath monitoring objective detects, and replaces the two-dimensional signal of plane with the three-dimensional information that contains distance, improvesThe precision detecting, and method based on computer learning, to the monitoring objective (example in visual fieldAs: people and vehicle etc.) carry out behavioural analysis, and right according to the result (identifying information) of analyzingClassification is carried out in behavior, and the behavior that exceedes certain rank is carried out to early warning.
Embodiment tri-
Based on embodiment mono-and embodiment bis-, embodiment tri-and embodiment tetra-provide a kind of based onDepth image is realized the application scenarios of road monitoring method, is intended to utilize computer vision technique pairThat road carries out is round-the-clock, monitoring in real time, and monitored results is analyzed, realize such as people orWagon flow monitoring and behavioural analysis, parking position monitoring, traffic violations monitoring, car plate identification, aerial andThe functions such as ground foreign matter monitoring (as falling object from high altitude, well lid).
Embodiment tri-is illustrated a kind of application scenarios, for example, and at preventing road monitoring systemAfter installation, while initialization, collecting unit can complete pavement detection automatically, identifies allAs lane line, dotted line, solid line, stop line, double amber lines, pavement, parking stall, well lid etc.The static mark in road surface and object, be stored in the above-mentioned image collecting initially as image templateChange in unit, then utilize RGB image and three-dimensional information to demarcate the position of the whole audience,And record the master sample of its primary data as follow-up data processing.
Parking stall and road occupying monitoring in violation of rules and regulations: parking stall has the parking bit line of standard, system conventionallyIn the time initializing detection, can automatically detect and identify parking bit line, and add up parking stall quantity. SystemSystem can be analyzed according to RGB image and three-dimensional information in real time, retrieves current parking stall residueSituation, in the time having vehicle to sail into or roll parking stall away from, system can identify and judge that parking stall accounts forBy situation, and upgrade residue parking stall counting, and be uploaded to judge module. Depth calculation simultaneouslyModule obtains three-dimensional information by RGB image calculation, then with image capture module in RGBImage is jointly uploaded to identification module and carries out the identification of image and the management of data. If roadside does not haveHave parking stall, and vehicle is while occurring that the long period is slack, system can be sentenced according to pre-conditionedWhether disconnected be road occupying in violation of rules and regulations, and report remote monitoring unit.
The present invention can be by the two-dimensional signal in two RGB collected by camera roads and three-dimensional letterBreath, has improved the precision detecting, and realizes multidirectional can be to information monitoring on road time againReport to the police, the real monitoring system that realizes intellectuality, automation.
Embodiment tetra-
The present embodiment is the road monitoring system based on depth image as embodiment mono-and embodiment bis-The another kind of application scenarios of system and method, for example, vehicle speed measuring: due to the position, track in visual fieldPut the RGB image and the three-dimensional information that have gathered by binocular camera and demarcate, vehicle sails intoTime, the position that can occur in successive video frames according to vehicle, by run duration with by the degree of depthThe translational speed of the move distance measuring vehicle that information obtains. Vehicle flowrate monitoring: utilize computerVision technique can measure in visual field vehicle and in the unit interval, enter visual field and leave and lookThe vehicle fleet size of field, thus calculate vehicle flowrate.
Aerial parabolic and the monitoring of ground foreign matter: by carrying out difference image fortune with the Background of standardCalculate, can extract the situation of aerial parabolic and ground foreign matter. And further combined with road imageThree-dimensional information, can obtain the three-dimensional geometry size of foreign matter, then according to the rule of design in advanceDetermine whether to charge center and send alarm signal. This system can be identified such as road surface unexpectedThere is sedimentation, the hazard events such as manhole cover loss.
In sum, comprehensively judge the two dimension letter on road by RGB image and depth imageBreath and three-dimensional information, improve accuracy and the reliability of monitoring; Based on computer vision technique,Road is carried out to target detection, based on the method for computer learning, to the monitoring in road imageTarget is carried out behavioural analysis, and according to identifying information, behavior is carried out to classification, to exceeding certain levelEarly warning is carried out in other behavior; Use computer vision technique to carry out in real time the content of road imageAnalysis, replace manual detection, judgement, avoid the erroneous judgement producing because of human factor.
By explanation and accompanying drawing, the typical case who has provided the ad hoc structure of detailed description of the invention implementsExample, based on spirit of the present invention, also can do other conversion. Although foregoing invention has proposed existingPreferred embodiment, but, these contents not as limitation.
For a person skilled in the art, read after above-mentioned explanation various changes and modificationsUndoubtedly will be apparent. Therefore, appending claims should regard as contain of the present invention trueWhole variations and the correction of sincere figure and scope. Any and all etc. within the scope of claimsThe scope of valency and content, all should think and still belong to the intent and scope of the invention.
Claims (22)
1. the preventing road monitoring system based on depth image, is characterized in that, described roadMonitoring system comprises:
Image capture module, comprises multiple collecting units, described multiple collecting unit synchronous acquisitionsRoad information is to obtain multiple road images;
Depth calculation module, is connected with described image capture module, to described multiple road imagesIn same place mate, and calculate three of each pixel in described multiple road imageDimension information;
Identification module, is connected with described image capture module, described depth calculation module respectively,The monitoring objective on road is carried out according to described multiple road images and described three-dimensional informationIdentify and then obtain identifying information;
Judge module, is connected with described identification module, in described judge module, is pre-stored with judgementRule, and described judge module judges described identifying information according to described judgment ruleAnd export judged result.
2. the preventing road monitoring system based on depth image according to claim 1, its spyLevy and be, described preventing road monitoring system comprises described far-end server and terminal device, wherein,
Described terminal device comprises described image capture module;
Described far-end server comprise described depth calculation module, described identification module and described in sentenceDisconnected module.
3. the preventing road monitoring system based on depth image according to claim 1, its spyLevy and be, described image capture module comprises multiple RGB cameras, described multiple RGB camerasTake respectively road image, to obtain RGB image.
4. the preventing road monitoring system based on depth image according to claim 1, its spyLevy and be, the setting of aliging in the horizontal direction of its projection centre of described multiple collecting units.
5. the preventing road monitoring system based on depth image according to claim 1, its spyLevy and be, described depth calculation module comprises:
Acquiring unit, is connected with described image capture module, obtains described image capture module and adoptsDescribed multiple road images of collection;
Demarcate unit, be connected with described acquiring unit, adopt computer vision technique to described manyIndividual collecting unit carries out respectively the demarcation of position, obtains multiple collecting unit coordinates;
Matching unit, is connected with described acquiring unit, of the same name in described multiple road imagesPoint mates, and extracts multiple characteristic points of described monitoring objective;
Computing unit, is connected with described demarcation unit, described matching unit respectively, according to describedMultiple collecting unit coordinate utilizations recovery formula calculate respectively the three-dimensional letter of described multiple characteristic pointsBreath, to obtain the three-dimensional information of described monitoring objective.
6. the preventing road monitoring system based on depth image according to claim 5, its spyLevy and be, described matching unit is by carrying out rim detection and characteristic matching to described monitoring objectiveMethod extract multiple characteristic points of described monitoring objective.
7. the preventing road monitoring system based on depth image according to claim 5, its spyLevy and be, described image capture module comprises the first collecting unit and the second collecting unit, described inDemarcate unit and demarcate the collecting unit coordinate P of described the first collecting unitaFor x1，y1, described markThe collecting unit coordinate P of described the second collecting unit demarcates in order unitbFor x2，y2, a characteristic pointThree-dimensional information be XC，YC，ZC, and the recovery formula of this characteristic point is:
Wherein, | Pa-Pb| be the road of described the first collecting unit and the collection of the second collecting unitThe position skew of image, the focal length that f is collecting unit, BCFor the throwing of described the first collecting unitThe line distance of the projection centre of shadow center and described the second collecting unit.
8. the preventing road monitoring system based on depth image according to claim 1, its spyLevy and be, described identification module comprises:
Pretreatment unit, detects monitoring objective in described road image and described monitoring objectiveThree-dimensional information;
Recognition unit, is connected with described detecting unit, by described monitoring objective and described three-dimensional letterBreath mates, identifies with image template, obtains described identifying information.
9. the preventing road monitoring system based on depth image according to claim 8, its spyLevy and be, described identification module also comprises:
Initialization unit, is connected with described recognition unit, is pre-stored with described image template.
10. the preventing road monitoring system based on depth image according to claim 1, its spyLevy and be, described identifying information comprises: the position of described monitoring objective and/or speed and/or threeDimension geometric scale.
11. preventing road monitoring systems based on depth image according to claim 1, its spyLevy and be, described preventing road monitoring system also comprises:
Memory module, respectively with described image capture module, described depth calculation module connects,Store described road image and described three-dimensional information.
12. preventing road monitoring systems based on depth image according to claim 1, its spyLevy and be, described judge module comprises:
Rule unit, is pre-stored with judgment rule;
Judging unit, judges described identifying information, obtains described judged result;
Whether output unit, is connected with described judging unit, determine defeated according to described judgment ruleGo out the control instruction of described judged result.
13. preventing road monitoring systems based on depth image according to claim 1, its spyLevy and be, described preventing road monitoring system also comprises:
Executive Module, is connected with described judge module, carries out described control instruction.
14. preventing road monitoring systems based on depth image according to claim 13, itsBe characterised in that, described Executive Module comprises:
Signal transmitter unit, is connected with described judge module, sends described control instruction;
Signal receiving unit, is connected with described signal transmitter unit, receives described control instruction;
Alarm unit, is connected with described signal receiving unit, according to the described control instruction receivingReport to the police;
Remote monitoring unit, is connected with described signal receiving unit, utilizes described control instruction to enterRow is further differentiated operation.
15. preventing road monitoring systems based on depth image according to claim 14, itsBe characterised in that, described alarm unit comprises:
Fixed terminal alarm unit, reports to the police in the position presetting;
Mobile terminal alarm unit, the mode of the movement of employing is reported to the police.
16. 1 kinds of road monitoring methods based on depth image, is characterized in that, described roadMethod for supervising comprises:
Step S1, multiple collecting unit synchronous acquisition road informations, obtain multiple road images;
Step S2, mates the same place in described multiple road images, and calculatesThe three-dimensional information of each pixel in described multiple road image;
Step S3, according to described multiple road images and described three-dimensional information to the prison on roadControl target is identified, and obtains identifying information;
Step S4, utilizes judgment rule to judge described identifying information, obtains judgement knotReally, and determine whether according to described judged result output control instruction.
17. road monitoring methods based on depth image according to claim 16, itsBe characterised in that, described road monitoring method also comprises:
Step S11, before described step S1, gathers standard road image, synthetic imageTemplate.
18. road monitoring methods based on depth image according to claim 16, itsBe characterised in that, described step S2 specifically comprises:
Step S21, obtains described multiple road images that described image capture module gathers;
Step S22, adopts computer vision technique to carry out respectively position to described multiple collecting unitsThe demarcation of putting, obtains multiple collecting unit coordinates;
Step S23, mates the same place in described multiple road images, described in extractionMultiple characteristic points of monitoring objective;
Step S24, recovers formula according to described multiple collecting unit coordinate utilizations and calculates respectively instituteState the three-dimensional information of multiple characteristic points, to obtain the three-dimensional information of described monitoring objective.
19. road monitoring methods based on depth image according to claim 16, itsBe characterised in that, described step S3 specifically comprises:
Step S31, detects three of monitoring objective in described road image and described monitoring objectiveDimension information;
Step S32, by described monitoring objective mate with described image template with described three-dimensional information,Identification, obtains described identifying information.
20. road monitoring methods based on depth image according to claim 16, itsBe characterised in that, step S4 specifically comprises:
Step S41, pre-stored judgment rule;
Step S42, judges described identifying information, obtains described judged result;
Step S43, determines the control of whether exporting described judged result according to described judgment ruleInstruction.
21. road monitoring methods based on depth image according to claim 16, itsBe characterised in that, described road monitoring method also comprises:
Step S5, after described step S4, carries out described control instruction.
22. road monitoring methods based on depth image according to claim 21, itsBe characterised in that, described step S5 specifically comprises:
Step S51, sends described control instruction;
Step S52, receives described control instruction;
Step S53, reports to the police according to the described control instruction receiving;
Step S54, utilizes described control instruction to carry out further differentiating operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610120682.9A CN105654732A (en) | 2016-03-03 | 2016-03-03 | Road monitoring system and method based on depth image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610120682.9A CN105654732A (en) | 2016-03-03 | 2016-03-03 | Road monitoring system and method based on depth image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105654732A true CN105654732A (en) | 2016-06-08 |
Family
ID=56493025
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610120682.9A Pending CN105654732A (en) | 2016-03-03 | 2016-03-03 | Road monitoring system and method based on depth image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105654732A (en) |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106448162A (en) * | 2016-09-30 | 2017-02-22 | 广东中星微电子有限公司 | Road monitoring method and road monitoring device |
| CN107085869A (en) * | 2017-01-16 | 2017-08-22 | 田桂昌 | Data handling system based on Internet of Things and cloud computing |
| CN107767687A (en) * | 2017-09-26 | 2018-03-06 | 中国科学院长春光学精密机械与物理研究所 | Free parking space detection method and system based on binocular stereo vision |
| CN107895386A (en) * | 2017-11-14 | 2018-04-10 | 中国航空工业集团公司西安飞机设计研究所 | A kind of multi-platform joint objective autonomous classification method |
| CN108121941A (en) * | 2016-11-30 | 2018-06-05 | 上海联合道路交通安全科学研究中心 | A kind of object speed calculation method based on monitoring device |
| CN108965651A (en) * | 2017-05-19 | 2018-12-07 | 深圳市道通智能航空技术有限公司 | A kind of drone height measurement method and unmanned plane |
| CN109117691A (en) * | 2017-06-23 | 2019-01-01 | 百度在线网络技术（北京）有限公司 | Drivable region detection method, device, equipment and storage medium |
| CN109191887A (en) * | 2018-08-06 | 2019-01-11 | 刘丽 | Localization method and positioning system in garage based on depth recognition |
| CN109308809A (en) * | 2017-07-29 | 2019-02-05 | 周晓辉 | A kind of tunnel device for monitoring vehicle based on dynamic image characteristic processing |
| CN109522840A (en) * | 2018-11-16 | 2019-03-26 | 孙睿 | A kind of expressway vehicle density monitoring calculation system and method |
| CN109886271A (en) * | 2019-01-22 | 2019-06-14 | 浙江大学 | It merges deep learning network and improves the image Accurate Segmentation method of edge detection |
| CN110099215A (en) * | 2019-05-06 | 2019-08-06 | 深圳市华芯技研科技有限公司 | A kind of method and apparatus extending binocular camera orientation range |
| CN110096975A (en) * | 2019-04-17 | 2019-08-06 | 北京筑梦园科技有限公司 | A kind of parking space state recognition methods, equipment and system |
| CN110378276A (en) * | 2019-07-16 | 2019-10-25 | 顺丰科技有限公司 | Vehicle-state acquisition methods, device, equipment and storage medium |
| CN110556000A (en) * | 2018-06-04 | 2019-12-10 | 义硕智能股份有限公司 | Artificial intelligence traffic detection system |
| CN110599739A (en) * | 2019-10-11 | 2019-12-20 | 宁夏广天夏电子科技有限公司 | Personnel safety detection system based on three-dimensional image and intelligent video technology |
| CN110672212A (en) * | 2019-09-19 | 2020-01-10 | 珠海格力电器股份有限公司 | Spatial three-dimensional temperature field detection method, computer readable storage medium and intelligent household appliance |
| CN110853367A (en) * | 2019-12-02 | 2020-02-28 | 江苏纬信工程咨询有限公司 | Intelligent risk prompt indicating system and method for road driving safety |
| CN110941991A (en) * | 2019-10-28 | 2020-03-31 | 成都华迈通信技术有限公司 | Multi-channel structured data acquisition system and data acquisition method |
| CN111785083A (en) * | 2020-05-27 | 2020-10-16 | 诠航科技有限公司 | Intelligent traffic early warning system based on guardrail patrol machine |
| CN111932635A (en) * | 2020-08-07 | 2020-11-13 | 江苏普达迪泰科技有限公司 | Image calibration method combining two-dimensional and three-dimensional visual processing |
| CN112488024A (en) * | 2020-12-08 | 2021-03-12 | 平安国际智慧城市科技股份有限公司 | High-altitude parabolic tracing method and device, computer equipment and storage medium |
| CN112581454A (en) * | 2020-12-23 | 2021-03-30 | 中原动力智能机器人有限公司 | Manhole cover missing detection method and system based on depth camera |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070053584A1 (en) * | 2001-11-09 | 2007-03-08 | Honda Giken Kogyo Kabushiki Kaisha | Image recognition apparatus |
| CN102183524A (en) * | 2011-01-10 | 2011-09-14 | 哈尔滨工业大学 | Double-CCD (Charge Coupled Device) detecting method and system for apparent defect assessment of civil engineering structure |
| CN104933718A (en) * | 2015-06-23 | 2015-09-23 | 广东省自动化研究所 | Physical coordinate positioning method based on binocular vision |
| CN105262983A (en) * | 2015-09-15 | 2016-01-20 | 上海图甲信息科技有限公司 | Road monitoring system and method based on internet of lamps |
-
2016
- 2016-03-03 CN CN201610120682.9A patent/CN105654732A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070053584A1 (en) * | 2001-11-09 | 2007-03-08 | Honda Giken Kogyo Kabushiki Kaisha | Image recognition apparatus |
| CN102183524A (en) * | 2011-01-10 | 2011-09-14 | 哈尔滨工业大学 | Double-CCD (Charge Coupled Device) detecting method and system for apparent defect assessment of civil engineering structure |
| CN104933718A (en) * | 2015-06-23 | 2015-09-23 | 广东省自动化研究所 | Physical coordinate positioning method based on binocular vision |
| CN105262983A (en) * | 2015-09-15 | 2016-01-20 | 上海图甲信息科技有限公司 | Road monitoring system and method based on internet of lamps |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106448162A (en) * | 2016-09-30 | 2017-02-22 | 广东中星微电子有限公司 | Road monitoring method and road monitoring device |
| CN108121941A (en) * | 2016-11-30 | 2018-06-05 | 上海联合道路交通安全科学研究中心 | A kind of object speed calculation method based on monitoring device |
| CN107085869A (en) * | 2017-01-16 | 2017-08-22 | 田桂昌 | Data handling system based on Internet of Things and cloud computing |
| CN108965651A (en) * | 2017-05-19 | 2018-12-07 | 深圳市道通智能航空技术有限公司 | A kind of drone height measurement method and unmanned plane |
| CN109117691A (en) * | 2017-06-23 | 2019-01-01 | 百度在线网络技术（北京）有限公司 | Drivable region detection method, device, equipment and storage medium |
| CN109308809A (en) * | 2017-07-29 | 2019-02-05 | 周晓辉 | A kind of tunnel device for monitoring vehicle based on dynamic image characteristic processing |
| CN107767687A (en) * | 2017-09-26 | 2018-03-06 | 中国科学院长春光学精密机械与物理研究所 | Free parking space detection method and system based on binocular stereo vision |
| CN107767687B (en) * | 2017-09-26 | 2020-09-01 | 中国科学院长春光学精密机械与物理研究所 | Free parking space detection method and system based on binocular stereoscopic vision |
| CN107895386A (en) * | 2017-11-14 | 2018-04-10 | 中国航空工业集团公司西安飞机设计研究所 | A kind of multi-platform joint objective autonomous classification method |
| CN110556000A (en) * | 2018-06-04 | 2019-12-10 | 义硕智能股份有限公司 | Artificial intelligence traffic detection system |
| CN109191887A (en) * | 2018-08-06 | 2019-01-11 | 刘丽 | Localization method and positioning system in garage based on depth recognition |
| CN109522840B (en) * | 2018-11-16 | 2023-05-30 | 孙睿 | Expressway vehicle flow density monitoring and calculating system and method |
| CN109522840A (en) * | 2018-11-16 | 2019-03-26 | 孙睿 | A kind of expressway vehicle density monitoring calculation system and method |
| CN109886271A (en) * | 2019-01-22 | 2019-06-14 | 浙江大学 | It merges deep learning network and improves the image Accurate Segmentation method of edge detection |
| CN110096975B (en) * | 2019-04-17 | 2021-04-09 | 北京筑梦园科技有限公司 | Parking space state identification method, equipment and system |
| CN110096975A (en) * | 2019-04-17 | 2019-08-06 | 北京筑梦园科技有限公司 | A kind of parking space state recognition methods, equipment and system |
| CN110099215A (en) * | 2019-05-06 | 2019-08-06 | 深圳市华芯技研科技有限公司 | A kind of method and apparatus extending binocular camera orientation range |
| CN110378276A (en) * | 2019-07-16 | 2019-10-25 | 顺丰科技有限公司 | Vehicle-state acquisition methods, device, equipment and storage medium |
| CN110378276B (en) * | 2019-07-16 | 2021-11-30 | 顺丰科技有限公司 | Vehicle state acquisition method, device, equipment and storage medium |
| CN110672212A (en) * | 2019-09-19 | 2020-01-10 | 珠海格力电器股份有限公司 | Spatial three-dimensional temperature field detection method, computer readable storage medium and intelligent household appliance |
| CN110599739A (en) * | 2019-10-11 | 2019-12-20 | 宁夏广天夏电子科技有限公司 | Personnel safety detection system based on three-dimensional image and intelligent video technology |
| CN110941991A (en) * | 2019-10-28 | 2020-03-31 | 成都华迈通信技术有限公司 | Multi-channel structured data acquisition system and data acquisition method |
| CN110941991B (en) * | 2019-10-28 | 2023-05-23 | 成都华迈通信技术有限公司 | Multichannel structured data acquisition system and data acquisition method |
| CN110853367A (en) * | 2019-12-02 | 2020-02-28 | 江苏纬信工程咨询有限公司 | Intelligent risk prompt indicating system and method for road driving safety |
| CN111785083A (en) * | 2020-05-27 | 2020-10-16 | 诠航科技有限公司 | Intelligent traffic early warning system based on guardrail patrol machine |
| CN111932635A (en) * | 2020-08-07 | 2020-11-13 | 江苏普达迪泰科技有限公司 | Image calibration method combining two-dimensional and three-dimensional visual processing |
| CN111932635B (en) * | 2020-08-07 | 2023-11-17 | 江苏普达迪泰科技有限公司 | Image calibration method adopting combination of two-dimensional and three-dimensional vision processing |
| CN112488024A (en) * | 2020-12-08 | 2021-03-12 | 平安国际智慧城市科技股份有限公司 | High-altitude parabolic tracing method and device, computer equipment and storage medium |
| CN112581454A (en) * | 2020-12-23 | 2021-03-30 | 中原动力智能机器人有限公司 | Manhole cover missing detection method and system based on depth camera |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105654732A (en) | Road monitoring system and method based on depth image | |
| CN106845547B (en) | A kind of intelligent automobile positioning and road markings identifying system and method based on camera | |
| CN102724482B (en) | Based on the intelligent vision sensing network moving target relay tracking system of GPS and GIS | |
| CN112700470B (en) | Target detection and track extraction method based on traffic video stream | |
| CN102435174B (en) | Method and device for detecting barrier based on hybrid binocular vision | |
| CN114359181B (en) | Intelligent traffic target fusion detection method and system based on image and point cloud | |
| CN100437660C (en) | Device for monitoring vehicle breaking regulation based on all-position visual sensor | |
| CN108229366A (en) | Deep learning vehicle-installed obstacle detection method based on radar and fusing image data | |
| CN111458721B (en) | Exposed garbage identification and positioning method, device and system | |
| CN108594244B (en) | Obstacle recognition transfer learning method based on stereoscopic vision and laser radar | |
| CN111275960A (en) | Traffic road condition analysis method, system and camera | |
| CN106019264A (en) | Binocular vision based UAV (Unmanned Aerial Vehicle) danger vehicle distance identifying system and method | |
| CN105608417A (en) | Traffic signal lamp detection method and device | |
| CN110415299B (en) | Vehicle position estimation method based on set guideboard under motion constraint | |
| CN114200442B (en) | Road target detection and association method based on millimeter wave radar and vision | |
| CN114463303B (en) | Road target detection method based on fusion of binocular camera and laser radar | |
| CN105262983A (en) | Road monitoring system and method based on internet of lamps | |
| CN106446785A (en) | Passable road detection method based on binocular vision | |
| CN114639115B (en) | Human body key point and laser radar fused 3D pedestrian detection method | |
| CN115004273A (en) | Digital reconstruction method, device and system for traffic road | |
| CN117111085A (en) | Automatic driving automobile road cloud fusion sensing method | |
| CN115440034B (en) | Vehicle-road cooperation realization method and realization system based on camera | |
| CN114067287A (en) | Foreign matter identification and early warning system based on vehicle side road side data perception fusion | |
| CN115690713A (en) | Binocular camera-based radar-vision fusion event detection method | |
| CN116912786A (en) | Intelligent network-connected automobile multi-mode fusion detection method based on vehicle-road cooperation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160608 |
|
| RJ01 | Rejection of invention patent application after publication |