CN107798302A - A kind of intelligent checking system and method for car mounting luggage frame - Google Patents
A kind of intelligent checking system and method for car mounting luggage frame Download PDFInfo
- Publication number
- CN107798302A CN107798302A CN201710949542.7A CN201710949542A CN107798302A CN 107798302 A CN107798302 A CN 107798302A CN 201710949542 A CN201710949542 A CN 201710949542A CN 107798302 A CN107798302 A CN 107798302A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- luggage carrier
- picture
- luggage
- detection
- 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.)
- Withdrawn
Links
Classifications
-
- 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
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention discloses a kind of intelligent checking system and method for car mounting luggage frame, including module of target detection and judge module, wherein, the module of target detection includes vehicle target detection unit, luggage carrier object detection unit and luggage carrier detection mark judging unit；The vehicle target detection unit obtains vehicle region image, the luggage carrier object detection unit is detected to vehicle region image and identifies luggage carrier, archives picture is compared luggage carrier detection mark judging unit with picture luggage carrier to be detected detection mark, and the determination module carries out synthetic determination to the result of whole testing process.Present invention is mainly applied to car mounting luggage frame in automotive vehicle annual test to detect, realize the whole-process automatic verification in detection process, unsanctioned detection image and reason can be passed back to server simultaneously preserve and remain to collect evidence, both save manpower, in turn ensure that the just, openly of verifying work.
Description
Technical field
The present invention relates to the artificial intelligence judgment technology field of automotive vehicle annual test, more particularly to a kind of car installs row additional
Lee's frame intelligent checking system and method.
Background technology
Constantly improve with living standards of the people with the continuous social and economic development, Urban vehicles poputation rapidly increases
It is long.Motor vehicle is as important traffic participant, it is necessary to possesses good safety and reliability.However, part motor vehicle is protected
The person's of having awareness of safety is thin, and refitted vehicles are carried out to motor vehicle.Vehicle after refitted vehicles, may without security test
Increase probability and the seriousness that traffic accident occurs.Therefore, strictly whether detection vehicle is reequiped for safeguarding that traffic safety is non-
It is often important.
Mainly by being accomplished manually, this method cost of labor is higher for traditional vehicle mounting luggage frame detection, efficiency compared with
It is low, and repeated verification operation easily produces fatigue for a long time, the defective mode such as easy carelessness, influences to verify accuracy rate.
How accurately and rapidly to vehicle, whether mounting luggage frame verifies, while avoids desk checking cost high, easily
Fatigue, the easily drawback such as carelessness, are to continue with the technical problem solved.
The content of the invention
The purpose of the present invention is：It is proposed a kind of car mounting luggage frame intelligent checking system and method, automatic detection vehicle
Whether mounting luggage frame, to meet nowadays the needs of to vehicle annual test operating efficiency, accuracy rate.
The technical solution adopted for the present invention to solve the technical problems is：
A kind of intelligent checking system of car mounting luggage frame, including module of target detection and judge module, wherein, it is described
Module of target detection includes vehicle target detection unit, luggage carrier object detection unit and luggage carrier detection mark judging unit；
The vehicle target detection unit detects vehicle image by vehicle target detection model, obtains vehicle region image, the row
Lee's frame object detection unit is detected using luggage carrier target detection model to vehicle region image, and identifies luggage carrier,
Archives picture is compared the luggage carrier detection mark judging unit with picture luggage carrier to be detected detection mark, described to sentence
Cover half block carries out synthetic determination to the result of whole testing process, and feeds back unsanctioned reason and picture.
A kind of intelligent detecting method of car mounting luggage frame, comprises the following steps：
S1, from server download vehicle pictures to be detected and map file picture；
S2, using based on deep learning network vehicle target detection model detect vehicle, judge vehicle pictures to be detected
Middle vehicle target whether there is, and is masked as 0 if recording this in the presence of if, extracts vehicle region image；This is recorded if being not present
Bar is masked as 1, and preserves picture concerned, into statistical analysis flow；
S3, using based on deep learning network vehicle target detection model detect vehicle, judge vehicle in archives picture
Target whether there is, and is masked as 0 if recording this in the presence of if, extracts vehicle region image；Indicate if recording this in the absence of if
For 1, and picture concerned is preserved, into statistical analysis flow；
S4, extracted from vehicle pictures to be checked using the luggage carrier target detection model inspection based on deep learning network
Vehicle region image, judge that luggage carrier whether there is, 0 is masked as if recording this in the presence of if, frame area image of picking up one's luggage；If
1 is masked as in the absence of this is then recorded；
S5, the vehicle extracted using the luggage carrier target detection model inspection based on deep learning network from archives picture
Area image, judge that luggage carrier whether there is, 0 is masked as if recording this in the presence of if, frame area image of picking up one's luggage；If do not deposit
1 is masked as then recording this；
S6, judge whether vehicle pictures luggage carrier detection mark to be detected and archives picture luggage carrier detection mark are consistent,
If consistent, record this and be masked as 0；If inconsistent, judge whether archives picture luggage carrier detection mark is 0, if 0
Record this and be masked as 0, if 1, then record this and be masked as 1；
S7, the result of the action to whole process carry out statistical analysis, if it is 0 to input this module identification, detection passes through；If
Mark 1 be present in the mark for inputting this module, then detect not by, while according to the position for being masked as 1 appearance can obtain detection not
Pass through reason and problem picture.
Further, the obtaining step of the vehicle target detection model is as follows：
S21, different automobile types are obtained in different illumination conditions, the vehicle image of different angle shooting；
S22, using rectangle frame marked vehicle area image position；
S23, deep neural network model, acquisition vehicle detection mould are detected using the vehicle region image training objective
Type.
Further, the obtaining step of the luggage carrier target detection model is as follows：
S31, the vehicle image for obtaining the different automobile types equipped with luggage carrier, vehicle body top luggage carrier region need complete；
S32, interception vehicle region image；
S33, luggage carrier position marked using rectangle frame；
S34, deep neural network model, acquisition luggage carrier detection are detected using the luggage carrier area image training objective
Model.
The beneficial effects of the invention are as follows：Present invention is mainly applied to car mounting luggage frame in automotive vehicle annual test to examine
Survey, realize the whole-process automatic verification in detection process, while unsanctioned detection image and reason can be passed back server
Preservation remains to collect evidence, and has both saved manpower, in turn ensure that the just, openly of verifying work.
Brief description of the drawings
Fig. 1 is the structural representation of the intelligent checking system of the present invention.
Fig. 2 is the detection decision flow chart of the mounting luggage frame of the present invention.
Fig. 3 is the structural representation of vehicle target detection unit of the present invention.
Fig. 4 is the structural representation of luggage carrier object detection unit of the present invention.
Embodiment
Below in conjunction with accompanying drawing.The present invention will be further described.
The intelligent checking system structure of the present invention is as shown in figure 1, including module of target detection and determination module.
Wherein, module of target detection includes：Vehicle target detection unit, luggage carrier object detection unit and luggage carrier detection
Indicate judging unit；
Vehicle target detection unit applies vehicle target detection model on vehicle image, obtains vehicle region image.So
Vehicle region image is passed to luggage carrier object detection unit afterwards, luggage carrier target detection mould is applied on vehicle region image
Type, identify luggage carrier.Module of target detection detects vehicle target first, and luggage carrier mesh is then detected in vehicle target image
Mark, this distribution detection means can be effectively prevented from because image background is complicated, in background comprising other baggage carrier of vehicle etc. because
Flase drop caused by element, improve the accuracy rate of luggage carrier detection.
Archives picture is compared luggage carrier detection mark judging unit with picture luggage carrier to be detected detection mark, sentences
Cover half block carries out synthetic determination to the result of whole testing process, and feeds back unsanctioned reason and picture.
The specific detection method of vehicle target detection unit includes：As shown in figure 3, detection module is first by vehicle to be detected
Image inputs vehicle target detection model, obtains N number of one-dimension array [class, x, y, width, height], array first first
Individual element represents object type, is that vehicle is then 1, is not that vehicle is then 0, four elements characterize square where destination object after array
Shape region, x, y represent rectangle upper left angular coordinate, and width represents rectangle width, and height represents rectangular elevation.Each array
A vehicle target is corresponded to, builds headlight for vehicle information using vehicle region rectangle frame size, with rectangle frame area most
Big array is exported as detection module, and vehicle region image is then extracted from image by rectangle frame positional information.This side
Method can effectively pick out other non-annual test target vehicles in background.
Vehicle target detection model acquisition methods are as follows：
S1, training data prepare：Acquisition different automobile types (vehicle such as such as car, sport car, offroad vehicle, minibus, commercial vehicle),
Different brands, the vehicle image for specifying shooting angle scope (being needed completely equipped with luggage carrier area image at the top of vehicle body)；
S2, data mark：Vehicle target is marked in the picture using rectangle frame, the corresponding rectangle frame of every image,
Inframe includes vehicle target；
S3, model training：Using the training data marked, the vehicle target based on deep learning network is trained to detect mould
Type (common knowledge, does not repeat hereby)；
The specific detection method of luggage carrier object detection unit includes：It is as shown in figure 4, obtained vehicle region image is defeated
Enter luggage carrier target detection model, obtain an one-dimension array [class, x, y, width, height], first element of array
Object type is represented, is that luggage carrier is then 1, is not that luggage carrier is then 0, four elements characterize rectangle where destination object after array
Region, x, y represent rectangle upper left angular coordinate, and width represents rectangle width, and height represents rectangular elevation, passes through rectangle frame
Positional information is picked up one's luggage frame area image from vehicle region image.
Luggage carrier target detection model acquisition methods are as follows：
S1, training data prepare：Obtain different automobile types, different brands, specify shooting angle scope (equipped with row at the top of vehicle body
Lee's frame area image needs complete) image, using the above-mentioned image of vehicle region target detection model batch processing, obtain vehicle area
Domain area image；
S2, data mark：Luggage carrier is marked in vehicle region image using rectangle frame, every vehicle region image pair
A rectangle frame is answered, inframe includes luggage carrier target；
S3, model training：Using the training data marked, the luggage carrier target detection based on deep learning network is trained
Model (common knowledge, does not repeat hereby)；
The mounting luggage frame examination criteria of the present invention is as follows：Vehicle target whether there is in picture to be detected；Archives picture
Interior vehicle target whether there is；Luggage carrier whether there is in picture vehicle region image to be detected；Archives picture vehicle region figure
As interior luggage carrier whether there is；Archives picture and picture luggage carrier to be detected detection mark comparative result；The present invention uses one
One-dimension array [x1, x2, x3, x4, x5] represents verification state, and initial value is [0,0,0,0,0].Flag bit x1 represents mapping to be checked
Vehicle target whether there is in piece, if in the presence of if x1 be 0, if in the absence of if x1 be 1；Flag bit x2 represents luggage in archives picture
Frame target whether there is, if in the presence of if x2 be 0, if in the absence of if x2 be 1；Flag bit x3 represents picture vehicle region figure to be detected
As interior luggage carrier whether there is, if in the presence of if x3 be 0, if in the absence of if x3 be 1；Flag bit x4 represents archives picture vehicle region
Luggage carrier whether there is in image, if in the presence of if x4 be 0, if in the absence of if x4 be 1；Flag bit x5 represent archives picture with it is to be checked
Mapping piece luggage carrier detection mark comparative result, if x3 is identical with x4 values, x5 0, if x3 is different from x4 values, and x4 is 0, then
X5 is 0, luggage carrier be present in this situation map file picture, and does not have luggage carrier in picture to be detected, is not belonging to increase checking
Scope；If x3 is different from x4 values, and x4 is 1, then x5 is 1, luggage carrier target be present in this situation map file picture to be detected,
And luggage carrier is not present in archives picture, belong to mounting luggage frame；Finally, statistical mark position [x1, x2, x5] state, if mark
To be 0, then verification passes through, if in the presence of 1, verifies and does not pass through.It can obtain verifying not passing through according to the position that state 1 occurs
The reason for.If x1 is 1, be not detected by vehicle target in image to be detected, it is possible the reason for have：Image to be detected obtains the stage
Error, vehicle shooting angle are against regulation, not bad comprising complete vehicle body or picture quality, overexposure or excessively dark occur, therefore
Examination ＆ verification is caused not pass through；If x2 is 1, vehicle target is not detected by archival image, possible cause is that file data obtains rank
Section error, or sorting phase makes a mistake during server storage archives picture, deposits other classification images by mistake, therefore cause examination ＆ verification not
Pass through.If x3 is 1, show the illegal mounting luggage frame of the vehicle, examination ＆ verification does not pass through.
Determination module according to verification standard judge luggage carrier verify whether by, if if direct back-checking successfully mark
Know, according to the position back-checking failure cause and corresponding picture that flag bit is 1, remain later stage examination ＆ verification if not and if investigate.
The implementation idiographic flow of the present invention is as shown in Fig. 2 comprise the following steps：
S1, from server download vehicle pictures to be detected and map file picture；
S2, using based on deep learning network vehicle target detection model detect vehicle, judge vehicle pictures to be detected
Middle vehicle target whether there is, and is masked as 0 if recording this in the presence of if, extracts vehicle region image；This is recorded if being not present
Bar is masked as 1, and preserves picture concerned, into statistical analysis flow；
S3, using based on deep learning network vehicle target detection model detect vehicle, judge vehicle in archives picture
Target whether there is, and is masked as 0 if recording this in the presence of if, extracts vehicle region image；Indicate if recording this in the absence of if
For 1, and picture concerned is preserved, into statistical analysis flow；
S4, extracted from vehicle pictures to be checked using the luggage carrier target detection model inspection based on deep learning network
Vehicle region image, judge that luggage carrier whether there is, 0 is masked as if recording this in the presence of if, frame area image of picking up one's luggage；If
1 is masked as in the absence of this is then recorded；
S5, the vehicle extracted using the luggage carrier target detection model inspection based on deep learning network from archives picture
Area image, judge that luggage carrier whether there is, 0 is masked as if recording this in the presence of if, frame area image of picking up one's luggage；If do not deposit
1 is masked as then recording this；
S6, judge whether vehicle pictures luggage carrier detection mark to be detected and archives picture luggage carrier detection mark are consistent,
If consistent, record this and be masked as 0；If inconsistent, judge whether archives picture luggage carrier detection mark is 0, if 0
Record this and be masked as 0, if 1, then record this and be masked as 1；
S7, the result of the action to whole process carry out statistical analysis, if it is 0 to input this module identification, detection passes through；If
Mark 1 be present in the mark for inputting this module, then detect not by, while according to the position for being masked as 1 appearance can obtain detection not
Pass through reason and problem picture.
The advantages of general principle and principal character and this programme of this programme has been shown and described above.The technology of the industry
Personnel are it should be appreciated that this programme is not restricted to the described embodiments, and the simply explanation described in above-described embodiment and specification is originally
The principle of scheme, on the premise of this programme spirit and scope are not departed from, this programme also has various changes and modifications, these changes
Change and improve and both fall within the range of claimed this programme.This programme be claimed scope by appended claims and its
Equivalent thereof.
Claims (4)
- A kind of 1. intelligent checking system of car mounting luggage frame, it is characterised in that including module of target detection and judge module, Wherein, the module of target detection includes vehicle target detection unit, luggage carrier object detection unit and luggage carrier detection mark Judging unit；The vehicle target detection unit detects vehicle image by vehicle target detection model, obtains vehicle region figure Picture, the luggage carrier object detection unit is detected using luggage carrier target detection model to vehicle region image, and is identified Go out luggage carrier, the luggage carrier detection mark judging unit is compared archives picture and picture luggage carrier to be detected detection mark Right, the determination module carries out synthetic determination to the result of whole testing process, and feeds back unsanctioned reason and picture.
- 2. a kind of intelligent detecting method of car mounting luggage frame, it is characterised in that comprise the following steps：S1, from server download vehicle pictures to be detected and map file picture；S2, using based on deep learning network vehicle target detection model detect vehicle, judge car in vehicle pictures to be detected Target whether there is, and is masked as 0 if recording this in the presence of if, extracts vehicle region image；Marked if recording this in the absence of if Will is 1, and preserves picture concerned, into statistical analysis flow；S3, using based on deep learning network vehicle target detection model detect vehicle, judge vehicle target in archives picture It whether there is, be masked as 0 if recording this in the presence of if, extract vehicle region image；1 is masked as if recording this in the absence of if, And picture concerned is preserved, into statistical analysis flow；S4, the vehicle extracted using the luggage carrier target detection model inspection based on deep learning network from vehicle pictures to be checked Area image, judge that luggage carrier whether there is, 0 is masked as if recording this in the presence of if, frame area image of picking up one's luggage；If do not deposit 1 is masked as then recording this；S5, the vehicle region extracted using the luggage carrier target detection model inspection based on deep learning network from archives picture Image, judge that luggage carrier whether there is, 0 is masked as if recording this in the presence of if, frame area image of picking up one's luggage；If in the absence of if Record this and be masked as 1；S6, judge whether vehicle pictures luggage carrier detection mark to be detected and archives picture luggage carrier detection mark are consistent, if one Cause, then record this and be masked as 0；If inconsistent, judge whether archives picture luggage carrier detection mark is 0, if 0 record This is masked as 0, if 1, then records this and is masked as 1；S7, the result of the action to whole process carry out statistical analysis, if it is 0 to input this module identification, detection passes through；If input Mark 1 be present in the mark of this module, then detect not by, while according to the position for being masked as 1 appearance can obtain detection do not pass through Reason and problem picture.
- 3. intelligent detecting method as claimed in claim 2, it is characterised in that the obtaining step of the vehicle target detection model It is as follows：S21, different automobile types are obtained in different illumination conditions, the vehicle image of different angle shooting；S22, using rectangle frame marked vehicle area image position；S23, deep neural network model, acquisition vehicle detection model are detected using the vehicle region image training objective.
- 4. intelligent detecting method as claimed in claim 2, it is characterised in that the acquisition step of the luggage carrier target detection model It is rapid as follows：S31, the vehicle image for obtaining the different automobile types equipped with luggage carrier, vehicle body top luggage carrier region need complete；S32, interception vehicle region image；S33, luggage carrier position marked using rectangle frame；S34, deep neural network model, acquisition luggage carrier detection mould are detected using the luggage carrier area image training objective Type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710949542.7A CN107798302A (en) | 2017-10-13 | 2017-10-13 | A kind of intelligent checking system and method for car mounting luggage frame |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710949542.7A CN107798302A (en) | 2017-10-13 | 2017-10-13 | A kind of intelligent checking system and method for car mounting luggage frame |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107798302A true CN107798302A (en) | 2018-03-13 |
Family
ID=61533088
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710949542.7A Withdrawn CN107798302A (en) | 2017-10-13 | 2017-10-13 | A kind of intelligent checking system and method for car mounting luggage frame |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107798302A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109637153A (en) * | 2019-01-25 | 2019-04-16 | 合肥市智信汽车科技有限公司 | A kind of vehicle-mounted mobile violation snap-shooting system based on machine vision |
| CN110728680A (en) * | 2019-10-25 | 2020-01-24 | 上海眼控科技股份有限公司 | Automobile data recorder detection method and device, computer equipment and storage medium |
| CN110751074A (en) * | 2019-10-12 | 2020-02-04 | 上海眼控科技股份有限公司 | Method, system and terminal for judging interior refitting of carriage and refitting judging method |
| WO2020029402A1 (en) * | 2018-08-08 | 2020-02-13 | 深圳码隆科技有限公司 | Suitcase volume measurement method and device, and suitcase volume measurement server, and computer readable medium |
-
2017
- 2017-10-13 CN CN201710949542.7A patent/CN107798302A/en not_active Withdrawn
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020029402A1 (en) * | 2018-08-08 | 2020-02-13 | 深圳码隆科技有限公司 | Suitcase volume measurement method and device, and suitcase volume measurement server, and computer readable medium |
| CN109637153A (en) * | 2019-01-25 | 2019-04-16 | 合肥市智信汽车科技有限公司 | A kind of vehicle-mounted mobile violation snap-shooting system based on machine vision |
| CN110751074A (en) * | 2019-10-12 | 2020-02-04 | 上海眼控科技股份有限公司 | Method, system and terminal for judging interior refitting of carriage and refitting judging method |
| CN110728680A (en) * | 2019-10-25 | 2020-01-24 | 上海眼控科技股份有限公司 | Automobile data recorder detection method and device, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107967445A (en) | A kind of car installs the intelligent checking system and method for skylight additional | |
| CN107958200A (en) | A kind of intelligent checking system and method for car repacking exhaust fan | |
| CN107798302A (en) | A kind of intelligent checking system and method for car mounting luggage frame | |
| CN107545239B (en) | Fake plate detection method based on license plate recognition and vehicle characteristic matching | |
| CN103903005B (en) | License plate image identification system and method | |
| CN107818322A (en) | A kind of vehicle VIN code tampering detection system and methods for vehicle annual test | |
| CN107958201A (en) | A kind of intelligent checking system and method for vehicle annual test insurance policy form | |
| CN111754456B (en) | Two-dimensional PCB appearance defect real-time automatic detection technology based on deep learning | |
| CN109784326A (en) | A kind of vehicle chassis detection method based on deep learning | |
| CN111126224A (en) | Vehicle detection method and classification recognition model training method | |
| CN106875697B (en) | Radio frequency identification and video identification comparison method and system | |
| CN104408475B (en) | A kind of licence plate recognition method and car license recognition equipment | |
| CN109741314A (en) | A kind of visible detection method and system of part | |
| CN109993138A (en) | A kind of car plate detection and recognition methods and device | |
| CN109344886B (en) | Occlusion number plate distinguishing method based on convolutional neural network | |
| CN109344835A (en) | Altering detecting method based on vehicle VIN code character position | |
| CN110378258B (en) | Image-based vehicle seat information detection method and device | |
| CN111080691A (en) | Infrared hot spot detection method and device for photovoltaic module | |
| CN107798301A (en) | A kind of signature detection system and method for vehicle annual test | |
| CN110288006A (en) | A kind of license plate number automatic Verification method and system | |
| CN103914682A (en) | Vehicle license plate recognition method and system | |
| CN105259304A (en) | On-line monitoring system and method for pollutants in vehicle tail gas | |
| CN107967446A (en) | A kind of intelligent checking system and method for installing engine protection device additional | |
| CN115810134B (en) | Image acquisition quality inspection method, system and device for vehicle insurance anti-fraud | |
| CN105989600A (en) | Characteristic point distribution statistics-based power distribution network device appearance detection method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20180313 |