CN103000067A - Right angle turning detection system and detecting method thereof - Google Patents
Right angle turning detection system and detecting method thereof Download PDFInfo
- Publication number
- CN103000067A CN103000067A CN2012105843476A CN201210584347A CN103000067A CN 103000067 A CN103000067 A CN 103000067A CN 2012105843476 A CN2012105843476 A CN 2012105843476A CN 201210584347 A CN201210584347 A CN 201210584347A CN 103000067 A CN103000067 A CN 103000067A
- Authority
- CN
- China
- Prior art keywords
- image
- camera
- coordinate
- angled bend
- point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a right angle turning detection system and detecting method thereof. The right angle turning detection system comprises a signboard, a camera, an image analyzer, a transmission network, a server and a client. The detecting method comprises that first recording the camera image of the right angle turning, measuring the right angle bend and the physical dimension of a training car through manual work and finishing the calibration of the coordinate system; the image analyzer continuously receive the monitoring images trapped by the camera and recognize the signboard on the training car, calculating the actual position, state, outline scope and the shortest distance from the outline to the marker line through the calibrated coordinated system, and capture images when the outline of the car crosses the lane line; the sever receive the measuring data and video come from the image analyze through the transmission network and calculate the training score, the driver and the trainer can query the information or read the video record through the rear client, thereby achieving a effective training on the driver and improving the learning efficiency of the drivers.
Description
Technical field
The present invention relates to a kind of right-angled bend detection system and detection method thereof.
Background technology
As automobile driver, right-angled bend is one of the driving efficiency that must skillfully grasp.But during traditional driver training, the student generally trains with fixed reference, and causing needs again to cultivate position sense and turning skill after actual the setting out on a journey.Although can be by the sensor collection vehicle position of some be installed on the road surface, the permanent magnet that adopts such as Chinese utility model 201120277850.8, or infrared ray commonly used, laser, ultrasound wave etc. are gone up in the engineering range finding, but sensors with auxiliary electrode were all can only perception on indivedual point of fixity positions, construction, maintenance complexity, overall cost is also expensive, and still needs linkage camera could realize evidence obtaining.
No matter the student learns or trainer's teaching, mainly all is based on visual information and judges, and it also is the easiest to be received therefore based on machine vision the whole dynamic process of right-angled bend being carried out detection mode.
Although China's utility model 201120489062.5 has been introduced a kind of video detecting device of the straight line distance for examination of driver, but its core is: " graphics processing unit is mapped to respectively HSV and LAB color space with the image that obtains; and in each spatial threshold value; form the straight line pixel; the pixel comparing unit uses principle of probability; the input pixel is added up, draw the straight line unit that meets the right side bearing feature most, calculate the rightmargin of this straight line unit ".Total institute is known, this mode is difficult to possess operability under actual environment, because there are many straight line pixels in vehicle self, such as vehicle window, ceiling etc., also often there are a large amount of straight line pixels in background environment simultaneously, such as kerbstone, flowers, plants and trees, lane line etc., in this case, be difficult in the pixel aspect " car " be extracted from background environment, finally cause detecting inaccurate.
Summary of the invention
The object of the invention is to provide a kind of right-angled bend parking detection system and detection method thereof, and it can realize training cart is carried out quantitative test in the right-angled bend process, is conducive to Real-Time Monitoring and the tracking of training result.
In order to solve these problems of the prior art, technical scheme provided by the invention is:
A kind of right-angled bend detection system, it comprises the back-end system that is installed near the front end system in right-angled bend position and is deployed in administrative center, described front end system comprises two main cameras, image analyzer and place sign board on the training cart, back-end system comprises server and client, front end system connects back-end system by transmission network, monitored picture covers the main camera in warehouse compartment zone for the travel conditions of shooting with video-corder warehouse compartment place training cart, the monitoring image that the lasting reception of the image analyzer that links to each other with main camera main camera is caught also is compressed into monitoring image, the operation attitude of while image analyzer monitor vehicle, image analyzer is sent to the server of back-end system with monitoring result and monitoring image, and client can be called monitoring result and the monitoring image of checking in the server.
As optimization, described two main cameras are fixed in the vertical rod of top, right-angled bend position, and each can overlook one section linearity region of quarter bend.
As optimization, also be provided with auxiliary camera in the both sides of every main camera, be used for the auxiliary running orbit that catches training cart in the warehouse compartment that replenishes.
As optimization, described notice plate is arranged on roof, engine machine cap or the case cover, and the marker graphic on the described notice plate is some gauge points, and gauge point is arranged in the specific line segment with fixed angle and unique intersection point.
As optimization, the seat in the plane of main camera, auxiliary camera is other to be equipped with light compensating lamp, and the synchronizing signal of described light compensating lamp is provided by main camera or auxiliary camera, is used for crossing in illumination carrying out the strobe type light filling when dark.
The present invention also provides a kind of right-angled bend detection method, it is characterized in that, described method relates to system and comprises the back-end system that is installed near the front end system in right-angled bend position and is deployed in administrative center, described front end system comprises many main cameras, auxiliary camera, image analyzer and place sign board on the training cart, back-end system comprises server and client, front end system connects back-end system by transmission network, monitored picture covers the main camera in warehouse compartment zone for the travel conditions of shooting with video-corder warehouse compartment place training cart, image analyzer and main camera, auxiliary camera links to each other, and concrete detection method may further comprise the steps:
Step S1: measure the quarter bend size by the direct labor, operation camera calibration program is finished image coordinate system to the conversion of world coordinate system;
Step S2: image analyzer continues to receive the right-angled bend monitoring image that main camera is caught, and is compressed into video record, simultaneously to monitoring image running mark board recognizer, the position that obtains gauge point and sign board;
Step S3: image analyzer is according to camera calibration result and sign board recognition result, and operational vehicle position and attitude trace routine obtains position and the attitude of vehicle; In conjunction with known vehicle dimension, and then definite vehicle ' s contour scope and profile outer edge are to the bee-line of quarter bend mark line; When the sign board of image analyzer identification disappears from the specific region of image, judge that vehicle sails out of corresponding main camera overlay area, image analyzer is metrical informations such as segmentation times spent, together with capture picture, video record all is sent to server;
Step S4: after server was received the metrical information that each image analyzer sends, the operation scoring procedures obtained this right-angled bend comprehensive grading;
Step S5: student or trainer browse, download and print the right-angled bend relevant scoring, capture picture and video record by the client-access server.
As optimization, among the step S3, trigger main camera 1 or auxiliary camera 2 candid photograph pictures when vehicle line ball or distance are less than setting threshold in docking process and be sent to image analyzer.
As optimization, among the step S1, when carrying out camera calibration, the step of described camera calibration program is:
Video camera is captured piece image, and take any pixel as the image coordinate system initial point, hand labeled goes out the image coordinate of four lane boundary line endpoints;
The physical location point at marking image coordinate origin place, and as the world coordinate system initial point, measure the world coordinates of four end points of lane line in the actual place of quarter bend;
World coordinates and the pixel coordinate substitution image-world coordinates of four lane boundary line endpoints are transformed in the system of linear equations, find the solution image-world coordinates transformed matrix;
Certain image that point coordinate value substitution image-the world coordinates transformed matrix is known under any one world coordinate system-world coordinates is transformed system of linear equations, can calculate image coordinate corresponding to this point; Under the prerequisite of the known world coordinate height, the image that any one image coordinate substitution image-world coordinates transformed matrix is known-world coordinates transforms system of linear equations, can calculate world coordinates corresponding to this point; So far, finish camera calibration
The steps flow chart that carries out the vehicle location attitude detection in this detection method among the step S3 is:
The image of the image coordinate substitution camera calibration program of the reference points detection figure that obtains in the sign board recognizer-world coordinates is transformed in the system of linear equations, can obtain the world coordinates of reference points detection figure;
In world coordinate system, the reference points detection figure is linked to be the line segment that has intersection point according to the queueing discipline of sign board mark original tally dot pattern;
In world coordinate system, calculate the coordinate of intersection point and the angle of line segment, intersecting point coordinate is vehicle location, and the line segment angle is vehicle attitude.
With respect to scheme of the prior art, advantage of the present invention is:
1. the invention describes a kind of right-angled bend detection system and detection method thereof, system comprises sign board, video camera, image analyzer, transmission network, server and client side; Detection method: record first the right-angled bend camera views, the physical size of manual measurement quarter bend, training cart is finished coordinate system and is demarcated; The monitoring image that the lasting reception of image analyzer video camera is caught and the sign board on the recognition training car, extrapolate physical location, attitude, profile scope and the profile of vehicle to the bee-line of mark line by the coordinate system of having demarcated, and when vehicle ' s contour is crossed over lane line, capture picture; Server receives from the measurement data of image analyzer and picture, video by transmission network, and calculate training scoring, driver and trainer inquire about or have access to video record by the client of rear end to information, so realize the effective training to the driver, improve the learning efficiency of driving the training personnel;
2. the present invention does not need ground on the scene to bury other sensor underground fully based on video technique, and also not needing increases any electric apparatus mounted on vehicle, simple in structure, with low cost;
3. the present invention supports the picture of the abnormal conditions such as line ball is captured directly based on video technique, and evidence obtaining is convenient.
Description of drawings
The invention will be further described below in conjunction with drawings and Examples:
Fig. 1 is the system architecture synoptic diagram of the embodiment of the invention;
Fig. 2 is the sign board synoptic diagram of the embodiment of the invention;
Fig. 3 is the overall system workflow diagram of the embodiment of the invention;
Fig. 4 is the camera calibration program flow diagram of the embodiment of the invention;
Fig. 5 is the sign board recognizer process flow diagram of the embodiment of the invention;
Fig. 6 is the vehicle location attitude detection process flow diagram of the embodiment of the invention;
Fig. 7 is the scoring procedures process flow diagram of the embodiment of the invention.
Wherein: 1, main camera; 2, auxiliary camera; 3, light compensating lamp; 4, image analyzer; 5, sign board; 6, transmission network; 7, server; 8, client.
Embodiment
Embodiment:
Present embodiment has been described a kind of right-angled bend detection system and detection method thereof, system architecture as shown in Figure 1, as shown in Figure 1, system is comprised of the back-end system that is installed near the front end system the quarter bend and is deployed in administrative center.Front end system mainly contains: main camera 1, auxiliary camera 2, light compensating lamp 3, image analyzer 4 and place sign board 5 on the training cart; Back-end system mainly contains: server 7, client 8; The front and back ends system is by transmission network 6 interconnection.
Take into account simultaneously the saving lower deployment cost in order to obtain good monitoring effect, a main camera 1 is installed on the vertical rod transverse arm of quarter bend porch apart from about 6 meters of intersection point, quarter bend outer side edges boundary line, another main camera 1 is installed on the vertical rod transverse arm of quarter bend exit apart from about 8 meters of intersection point, quarter bend outer side edges boundary line, highly about 6 meters of two vertical rods; Two main cameras 1 all place center line top, track.Main camera 1 adopts the wide temperature design of technical grade, sensor adopts 5,000,000 CCD, resolution is greater than 2592x1936, frame per second was not less than for 8 frame/seconds, built-in gigabit ethernet interface is equipped with many moneys of 10mm ~ 35mm tight shot or zoom wide-angle lens simultaneously, can guarantee effectively like this that monitored picture covers one section of quarter bend at least under the above-mentioned mounting condition to be not less than 4 meters linearity regions, and do not have the blind area, guarantee that simultaneously sign board pixels across width is no less than 500 pixels.Auxiliary camera 2 can be installed on the main camera 1 same vertical rod transverse arm, particular location is 0.2-1 rice outside the boundary line, the left and right sides, track, auxiliary camera adopts the wide temperature design of technical grade, sensor adopts 2,000,000 CCD, resolution is greater than 1920x1080, and frame per second is not less than 8 frame/seconds, built-in gigabit ethernet interface, be equipped with simultaneously many moneys of 10mm ~ 25mm tight shot or zoom lens, can effectively guarantee like this can effective monitoring wheel line ball under the above-mentioned mounting condition.
Light compensating lamp 3 adopts strobe type LED light compensating lamps, and synchronizing signal is main light compensating lamp by what main camera 1 provided, and synchronizing signal is assisted light compensating lamp by crying of providing of auxiliary camera.Light compensating lamp 3 power and visible angle are decided according to the installation site.For avoiding vehicle body reflection to cause the video camera overexposure, the installation that should keep at a certain distance away of light compensating lamp 3 and its synchronous video camera of control generally is not less than 0.5 meter.When light compensating lamp 2 setting height(from bottom)s are 6 meters, when being 6 meters apart from parking stall central horizontal distance, visible angle is not less than 40 degree, and power is not less than 15W, can satisfy the demands with video camera interval 1 meter ampere dress.
Image analyzer 4 adopts the built-in industrial control computer, and the integrated heat-dissipating casing of its shell does not need radiator fan, effectively prevents and treats inner laying dust, improves system stability; When connecting the ccd video camera of 5,000,000 pixels, require the configuration dominant frequency to be not less than the CPU of 2.4GHz, the internal memory that is no less than 4GB, the hard disk that is no less than 32GB is as hardware interfaces such as external memory and gigabit Ethernet, RS232, and above-mentioned configuration guarantees that image analyzer 4 has enough calculating, storage and the communication resource to move all kinds of image processing algorithms and application program.
Server 7 can adopt tower server, selects dominant frequency greater than four core processors of 2.8GHz, 8MB buffer memory, and internal memory is not less than 4GB, to guarantee still to guarantee the system responses real-time when disposing many image analyzers 4.
Client 8 is that common PC gets final product, and can be equipped with the peripherals such as printer, IC-card card reader or fingerprint capturer.
Transmission network 6 adopts optical fiber or 3G cordless communication network when cross-over connection front and back ends system, and front and back ends system this locality gigabit Ethernets that adopt based on twisted-pair feeder more.
Main camera 1, auxiliary camera 2, image analyzer 4, server 7, client 8 realize exchanges data by transmission network 6.
As shown in Figure 2, sign board 5 synoptic diagram of the present invention are comprised of three parts: gauge point figure, driving school's code name character, training cart numbering character: gauge point figure necessary energy clear and definite indicating signboard position and direction, as adopt T-shaped sign, and its horizontal line segment is comprised of the black circle diagram shape of 5 white backgrounds, vertical line segment is comprised of 3 white with black circle diagram shapes, laterally line segment can be distinguished and be identified after extracting by figure binaryzation, connected domain at an easy rate with vertical line segment, and laterally there is vertical angle in line segment with vertical line segment; Driving school's code name character adopts two English alphabets, refers to such as " Hua Feng driving school " available " HF "; Training cart numbering character adopts three arabic numeral.
Overall system workflow diagram of the present invention as shown in Figure 3:
Step S1: measure the quarter bend size by the direct labor, operation camera calibration program realizes the mutual conversion of image coordinate system and world coordinate system;
Step S2: image analyzer 4 continues to receive the monitoring image that main cameras 1 are caught, and is compressed into video record, and running mark board recognizer simultaneously obtains position and the direction of sign board 5;
Step S3: image analyzer 4 is according to main camera 1 calibration result and sign board 5 recognition results, and operational vehicle position and attitude trace routine obtains position and attitude in the vehicle world coordinate system; According to vehicle location, attitude and known vehicle dimension, determine vehicle ' s contour scope and profile outer edge to the bee-line of lane markings line, and at the vehicle line ball or distance triggers main camera 1 during less than setting threshold or auxiliary camera 2 is captured pictures.
Step S4: image analyzer 4 is with the vehicle location that obtains among the step S3, attitude, range finding result and capture the data communication devices such as picture, video record and cross transmission network 6 and be sent to server 7, server 7 service data supervisory routines are analyzed various information and are provided the application datas such as comprehensive grading, project time spent, mode with data recording deposits database in, deposit picture, video in the specified file path, move simultaneously application services support, administrative client access.
Step S5: student's operated client 8, by transmission network 6 logon servers 7, browse the information such as the relevant data recording of training, picture, video, and according to demand printing, Download Info; Trainer's operated client 8 by transmission network 6 logon servers 7, is browsed teaching person's the information such as data recording, picture, video, and according to demand print, Download Info.
The purpose of camera calibration is the corresponding relation that calculates world coordinate system and image coordinate system, world coordinate system is the coordinate system of the real world at vehicle place, by 3 coordinate axis: X-axis, Y-axis, Z axis form, image coordinate system is the plane picture of shot by camera, and by 2 coordinate axis: U axle, V axle form.The coordinate of supposing certain point in world coordinate system X, Y, three coordinate axis of Z is (x
w, y
w, z
w), its corresponding point coordinate in image coordinate system is (u, v), then the corresponding relation of this two point coordinate can be expressed as system of linear equations:
This system of equations is called " image-world coordinates transforms system of linear equations ", and wherein α is the intermediate parameters of computation process, and (u, v) is the image pixel coordinate, (x
w, y
w, z
w) be world coordinates, (f
x, f
y) be respectively the focal length of X-axis and Y direction in the image coordinate system, (u
0, v
0) be the position of intersection point in image coordinate of camera optical axis and the plane of delineation.T=[t
x, t
y, t
z, 1] and be the mapping parameters of world coordinates initial point in image coordinate, R is orthogonal matrix, is defined as:
Metzler matrix is image-world coordinate system transformation matrix, has comprised all camera calibration parameters to be determined, and camera calibration is exactly the process of finding the solution Metzler matrix.We are by measuring world coordinates the value { (x on four summits that obtain the training site
Wi, y
Wi, z
Wi) | i=1 ..., 4} obtains the coordinate figure { (u of these 4 angle points in image in the image by manual markings
i, v
i) | i=1 ..., 4} can obtain Metzler matrix by the method for separating above-mentioned system of linear equations.
After calculating Metzler matrix, appoint to certain point coordinate value under the world coordinate system, can calculate image coordinate value corresponding to this point; Appoint to the coordinate of certain point under the image coordinate system and the height z of known these corresponding point in world coordinates
w, can calculate the world coordinates of corresponding point, so the present invention adopts camera calibration program flow diagram as shown in Figure 4:
Step S101: video camera is captured piece image, and take any pixel as initial point, hand labeled goes out the image coordinate { (u of four lane boundary line endpoints of this section quarter bend
i, v
i) | i=1 ..., 4};
Step S102: the physical location point at marking image coordinate origin place, and as the world coordinate system initial point, measure the world coordinates { (x of four end points of lane line in the actual place of quarter bend
Wi, y
Wi, z
Wi) | i=1 ..., 4};
Step S103: world coordinates and the image coordinate substitution image-world coordinates of four lane boundary line endpoints of quarter bend are transformed in the system of linear equations, find the solution image-world coordinates transformed matrix M;
Step S104: certain point coordinate value substitution M is known under any one world coordinate system image-world coordinates is transformed system of linear equations, can calculate the image coordinate of this point correspondence; Under the prerequisite of the known world coordinate height, the image that any one image coordinate substitution image-world coordinates transformed matrix is known-world coordinates transforms system of linear equations, can calculate world coordinates corresponding to this point; So far, finish camera calibration.
When adopting T-shaped sign board shown in Figure 2, black circle is the Main Basis of distinguishing mark with white circle, and the present invention can adopt sign board recognizer process flow diagram as shown in Figure 5:
Step S201: establish the pixel value that image coordinate (x, y) locates and be designated as p
X, y, defining two width of cloth Threshold segmentation image LB, LD and be respectively applied to detect black circle and white circle, its pixel value is for being respectively:
T_B wherein, T_D is preset value.Calculate all pixel values of LB, LD according to following formula.
Step S202: be 1 connected region for detecting white circle, finding out among the LB, the length and width value that this is regional is converted to the length and width value under the world coordinate system, judges that the white circle whether the length and width value meet in the sign board 5 is big or small, meets then to be designated as the white circle that detects.In like manner, do same processing for LD, can detect black circle.
Step S203: utilize detected white circle and black circle, judge whether one by one to connect into the specific line segment consistent with sign board, if can, then look this white round, black circle and be the reference points detection figure, otherwise give up this white round or black circle; Repeat this step, until travel through all white circles and black circle;
Step S204: if step S203 finds the reference points detection figure, then take the reference points detection figure as reference point, locate other additional letter and numeric area, identification as for letter and number, can adopt the method identifications such as template matches, because character and numeral identification have been the common technologies that image is processed, in this patent and the non-key technologies point, do not repeat them here simultaneously.So far finish the identification of whole sign board 5.
Because sign board 5 is fixed position and the directions that are pasted on vehicle, so position and the attitude that can extrapolate vehicle according to position and the direction of sign board 5.Here suppose the dimensional parameters of vehicle, position and the direction of for example length, and sign board on vehicle body obtains by measuring.
When adopting T-shaped sign board 5 shown in Figure 2 and sign board detection method shown in Figure 3, the angle Theta of the line segments that the line segments that a plurality of black circles form and a plurality of white circle form is sign board 5 directions.The image coordinate of supposing sign is (u
i, v
i), angle is Theta, sign board 5 is pasted on roof, then z
w=height of car is known, so the position of positioned vehicle and attitude can adopt vehicle location attitude detection flow process as shown in Figure 6:
Step S301: with the image coordinate (u of the reference points detection figure that obtains in the sign board recognizer
i, v
i) image-world coordinates of substitution camera calibration program transforms in the system of linear equations, can obtain the world coordinates (x of reference points detection figure
w, y
w, z
w);
Step S302: in world coordinate system, the reference points detection figure is linked to be the line segment that has intersection point according to the queueing discipline of the original note dot pattern of sign board mark;
Step S303: in world coordinate system, calculate the coordinate of intersection point and the angle Theta of line segment
w, intersecting point coordinate is vehicle location, and the line segment angle is vehicle attitude.
Scoring procedures process flow diagram of the present invention as shown in Figure 7.
Step S401: whether statistics exists vehicle line ball, time spent to surpass the obviously misoperationes such as maximum time limit, if exist, judges directly that then this right-angled bend is defective, and scoring finishes;
Step S402: judge that the time spent is whether in critical field, if the critical field of being no more than, then do not deduct points, if surpass, then according to setting rule deduction corresponding scores, if the overtime upper limit that reaches, judge directly that then this right-angled bend is defective, for example detained 1 minute in every overtime 5 seconds, until overtime 200 seconds, namely the described time spent of step S401 surpasses the maximum time limit;
Step S403: add up final integrate score, if be lower than setting value, then judge defectively, otherwise be judged to be qualified and provide the judgement mark.
Above-mentioned example only is explanation technical conceive of the present invention and characteristics, and its purpose is to allow the people who is familiar with technique can understand content of the present invention and according to this enforcement, can not limit protection scope of the present invention with this.All equivalent transformations that Spirit Essence is done according to the present invention or modification all should be encompassed within protection scope of the present invention.
Claims (10)
1. right-angled bend detection system, it is characterized in that, it comprises the back-end system that is installed near the front end system in right-angled bend position and is deployed in administrative center, described front end system comprises two main cameras, image analyzer and place sign board on the training cart, back-end system comprises server and client, front end system connects back-end system by transmission network, monitored picture covers the main camera in warehouse compartment zone for the travel conditions of shooting with video-corder warehouse compartment place training cart, the monitoring image that the lasting reception of the image analyzer that links to each other with main camera main camera is caught also is compressed into monitoring image, the operation attitude of while image analyzer monitor vehicle, image analyzer is sent to the server of back-end system with monitoring result and monitoring image, and client can be called monitoring result and the monitoring image of checking in the server.
2. right-angled bend detection system according to claim 1 is characterized in that, described two main cameras are fixed in the vertical rod of top, right-angled bend position, and each can overlook one section linearity region of quarter bend.
3. right-angled bend detection system according to claim 2 is characterized in that, also is provided with auxiliary camera in the both sides of every main camera, is used for the auxiliary running orbit that catches training cart in the warehouse compartment that replenishes.
4. right-angled bend detection system according to claim 1, it is characterized in that, described notice plate is arranged on roof, engine machine cap or the case cover, and the marker graphic on the described notice plate is some gauge points, and gauge point is arranged in the specific line segment with fixed angle and unique intersection point.
5. according to claim 1 and 2 or 3 described right-angled bend detection systems, it is characterized in that, the seat in the plane of main camera, auxiliary camera is other to be equipped with light compensating lamp, and the synchronizing signal of described light compensating lamp is provided by main camera or auxiliary camera, is used for crossing in illumination carrying out the strobe type light filling when dark.
6. right-angled bend detection method, it is characterized in that, described method relates to system and comprises the back-end system that is installed near the front end system in right-angled bend position and is deployed in administrative center, described front end system comprises many main cameras, auxiliary camera, image analyzer and place sign board on the training cart, back-end system comprises server and client, front end system connects back-end system by transmission network, monitored picture covers the main camera in warehouse compartment zone for the travel conditions of shooting with video-corder warehouse compartment place training cart, image analyzer and main camera, auxiliary camera links to each other, and concrete detection method may further comprise the steps:
Step S1: measure the quarter bend size by the direct labor, operation camera calibration program is finished image coordinate system to the conversion of world coordinate system;
Step S2: image analyzer continues to receive the right-angled bend monitoring image that main camera is caught, and is compressed into video record, simultaneously to monitoring image running mark board recognizer, the position that obtains gauge point and sign board;
Step S3: image analyzer is according to camera calibration result and sign board recognition result, and operational vehicle position and attitude trace routine obtains position and the attitude of vehicle; In conjunction with known vehicle dimension, and then definite vehicle ' s contour scope and profile outer edge are to the bee-line of quarter bend mark line; When the sign board of image analyzer identification disappears from the specific region of image, judge that vehicle sails out of corresponding main camera overlay area, image analyzer is metrical informations such as segmentation times spent, together with capture picture, video record all is sent to server;
Step S4: after server was received the metrical information that each image analyzer sends, the operation scoring procedures obtained this right-angled bend comprehensive grading;
Step S5: student or trainer browse, download and print the right-angled bend relevant scoring, capture picture and video record by the client-access server.
7. right-angled bend detection method according to claim 6 is characterized in that, among the step S3, triggers main camera 1 or auxiliary camera 2 candid photograph pictures when vehicle line ball or distance are less than setting threshold in docking process and is sent to image analyzer.
8. right-angled bend detection method according to claim 6 is characterized in that, among the step S1, when carrying out camera calibration, the step of described camera calibration program is:
Video camera is captured piece image, and take any pixel as the image coordinate system initial point, hand labeled goes out the image coordinate of four lane boundary line endpoints;
The physical location point at marking image coordinate origin place, and as the world coordinate system initial point, measure the world coordinates of four end points of lane line in the actual place of quarter bend;
World coordinates and the pixel coordinate substitution image-world coordinates of four lane boundary line endpoints are transformed in the system of linear equations, find the solution image-world coordinates transformed matrix;
Certain image that point coordinate value substitution image-the world coordinates transformed matrix is known under any one world coordinate system-world coordinates is transformed system of linear equations, can calculate image coordinate corresponding to this point; Under the prerequisite of the known world coordinate height, the image that any one image coordinate substitution image-world coordinates transformed matrix is known-world coordinates transforms system of linear equations, can calculate world coordinates corresponding to this point; So far, finish camera calibration.
9. according to claim 6 or 7 or 8 described right-angled bend detection methods, it is characterized in that the flow process of carrying out sign board identification among the step S2 is as follows:
Image is carried out Threshold segmentation, the image after obtaining to cut apart;
Image behind the Threshold segmentation is carried out connected domain extract, judge whether the connected domain length and width meets the length and width setting value of sign board gauge point figure, if meet, this connected domain of mark then;
Can the connected domain of judge mark connect into the shape consistent with the real marking board one by one, comprises the angle that gauge point quantity becomes with the gauge point line, if can, then looking connected region is the reference points detection figure, otherwise gives up this connected domain; Repeat this step, until travel through all connected domains;
Take the reference points detection figure as reference point, identify other additional letter and numeric area and identification, so far finish the identification of whole sign board.
10. described right-angled bend detection method according to claim 9 is characterized in that, the steps flow chart that carries out the vehicle location attitude detection in this detection method among the step S3 is:
The image of the image coordinate substitution camera calibration program of the reference points detection figure that obtains in the sign board recognizer-world coordinates is transformed in the system of linear equations, can obtain the world coordinates of reference points detection figure;
In world coordinate system, the reference points detection figure is linked to be the line segment that has intersection point according to the queueing discipline of sign board mark original tally dot pattern;
In world coordinate system, calculate the coordinate of intersection point and the angle of line segment, intersecting point coordinate is vehicle location, and the line segment angle is vehicle attitude.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210584347.6A CN103000067B (en) | 2012-12-28 | 2012-12-28 | Right angle turning detection system and detecting method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210584347.6A CN103000067B (en) | 2012-12-28 | 2012-12-28 | Right angle turning detection system and detecting method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103000067A true CN103000067A (en) | 2013-03-27 |
| CN103000067B CN103000067B (en) | 2014-12-17 |
Family
ID=47928582
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210584347.6A Active CN103000067B (en) | 2012-12-28 | 2012-12-28 | Right angle turning detection system and detecting method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103000067B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105741643A (en) * | 2016-04-20 | 2016-07-06 | 北京百度网讯科技有限公司 | Method, device and system for realizing automatic driving skill training |
| CN106952308A (en) * | 2017-04-01 | 2017-07-14 | 上海蔚来汽车有限公司 | The location determining method and system of moving object |
| CN108364542A (en) * | 2018-01-12 | 2018-08-03 | 重庆韦娜软件有限公司 | Fork truck field test position detection and decision maker and its method |
| CN108550225A (en) * | 2018-04-09 | 2018-09-18 | 廖辉 | A kind of share examines vehicle system and the shared business model based on the system |
| CN109272821A (en) * | 2018-10-22 | 2019-01-25 | 广州星唯信息科技有限公司 | A kind of place driving evaluation method based on high-precision vision positioning |
| CN109813229A (en) * | 2018-12-30 | 2019-05-28 | 广州展讯信息科技有限公司 | Detection method and device of the vehicle body to sideline distance |
| CN112270873A (en) * | 2020-09-28 | 2021-01-26 | 中电海康集团有限公司 | Subject two-driving training teaching system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2600787Y (en) * | 2002-08-21 | 2004-01-21 | 张拥军 | Device for measuring running trace of vehicle starting using digital image identification technology |
| JP2005258364A (en) * | 2004-02-13 | 2005-09-22 | Toshiba Corp | Remote test and course system |
| CN201037962Y (en) * | 2006-06-29 | 2008-03-19 | 北京艾曼特科技有限公司 | Electronic pile examining system |
| CN102175172A (en) * | 2010-12-30 | 2011-09-07 | 南京理工大学 | Image-identification-based system for detecting contour dimension of motor vehicle |
-
2012
- 2012-12-28 CN CN201210584347.6A patent/CN103000067B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN2600787Y (en) * | 2002-08-21 | 2004-01-21 | 张拥军 | Device for measuring running trace of vehicle starting using digital image identification technology |
| JP2005258364A (en) * | 2004-02-13 | 2005-09-22 | Toshiba Corp | Remote test and course system |
| CN201037962Y (en) * | 2006-06-29 | 2008-03-19 | 北京艾曼特科技有限公司 | Electronic pile examining system |
| CN102175172A (en) * | 2010-12-30 | 2011-09-07 | 南京理工大学 | Image-identification-based system for detecting contour dimension of motor vehicle |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105741643A (en) * | 2016-04-20 | 2016-07-06 | 北京百度网讯科技有限公司 | Method, device and system for realizing automatic driving skill training |
| CN105741643B (en) * | 2016-04-20 | 2020-05-22 | 北京百度网讯科技有限公司 | Automatic implementation method, device and system for training driving technology |
| CN106952308A (en) * | 2017-04-01 | 2017-07-14 | 上海蔚来汽车有限公司 | The location determining method and system of moving object |
| CN106952308B (en) * | 2017-04-01 | 2020-02-28 | 上海蔚来汽车有限公司 | Method and system for determining position of moving object |
| CN108364542A (en) * | 2018-01-12 | 2018-08-03 | 重庆韦娜软件有限公司 | Fork truck field test position detection and decision maker and its method |
| CN108550225A (en) * | 2018-04-09 | 2018-09-18 | 廖辉 | A kind of share examines vehicle system and the shared business model based on the system |
| CN109272821A (en) * | 2018-10-22 | 2019-01-25 | 广州星唯信息科技有限公司 | A kind of place driving evaluation method based on high-precision vision positioning |
| CN109813229A (en) * | 2018-12-30 | 2019-05-28 | 广州展讯信息科技有限公司 | Detection method and device of the vehicle body to sideline distance |
| CN112270873A (en) * | 2020-09-28 | 2021-01-26 | 中电海康集团有限公司 | Subject two-driving training teaching system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103000067B (en) | 2014-12-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103065520B (en) | Detection system for backing car into storage and detection method thereof | |
| CN103065519B (en) | Detecting system of ramp fixed point stopping and starting and detecting method of ramp fixed point stopping and starting | |
| CN103076007B (en) | Lateral parking detection system and detection method thereof | |
| CN103000067B (en) | Right angle turning detection system and detecting method thereof | |
| CN106919915B (en) | Map road marking and road quality acquisition device and method based on ADAS system | |
| CN102867417B (en) | Taxi anti-forgery system and taxi anti-forgery method | |
| CN104848851B (en) | Intelligent Mobile Robot and its method based on Fusion composition | |
| CN110619279B (en) | Road traffic sign instance segmentation method based on tracking | |
| CN102073846B (en) | Method for acquiring traffic information based on aerial images | |
| US20130229517A1 (en) | Vehicle speed measurement method and system utilizing a single image capturing unit | |
| CN103630496B (en) | Based on the traffic video visibility detecting method of road surface apparent brightness and least square method | |
| CN104766086A (en) | Supervising method and system of way mark | |
| CN111339905B (en) | CIM well lid state visual detection system based on deep learning and multiple visual angles | |
| KR102119033B1 (en) | Apparatus and method for providing road marking information using image data and vibration | |
| CN106251695A (en) | Destination's parking stall intelligent recommendation system and method based on parking space state monitoring | |
| CN102446355B (en) | Method for detecting target protruding from plane based on double viewing fields without calibration | |
| CN103077639B (en) | Curve driving detection system and detection method thereof | |
| CN112735253B (en) | Traffic light automatic labeling method and computer equipment | |
| CN114812403A (en) | Large-span steel structure hoisting deformation monitoring method based on unmanned aerial vehicle and machine vision | |
| CN105869413A (en) | Method for measuring traffic flow and speed based on camera video | |
| CN114067287A (en) | Foreign matter identification and early warning system based on vehicle side road side data perception fusion | |
| CN110415299B (en) | Vehicle position estimation method based on set guideboard under motion constraint | |
| KR20210044127A (en) | Visual range measurement and alarm system based on video analysis and method thereof | |
| CN202887450U (en) | Taxi anti-fake system | |
| CN113177508B (en) | Method, device and equipment for processing driving information |
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 |