CN108805007A - Class's line hole capital after selling all securities state judging method based on image recognition - Google Patents
Class's line hole capital after selling all securities state judging method based on image recognition Download PDFInfo
- Publication number
- CN108805007A CN108805007A CN201810346729.2A CN201810346729A CN108805007A CN 108805007 A CN108805007 A CN 108805007A CN 201810346729 A CN201810346729 A CN 201810346729A CN 108805007 A CN108805007 A CN 108805007A
- Authority
- CN
- China
- Prior art keywords
- regular bus
- class
- selling
- hole capital
- photo
- 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
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
- G06Q10/083—Shipping
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Data Mining & Analysis (AREA)
- Quality & Reliability (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses a kind of class's line hole capital after selling all securities state judging method based on image recognition, including：Regular bus loading condition take pictures and obtains the regular bus photo；The corresponding job order of the regular bus is scanned and obtains the regular bus information；Class's line information that the regular bus corresponds to class's line is extracted in the regular bus information, the regular bus photo is associated with class's line information；The regular bus photo is identified based on deep neural network, obtains recognition result；Associated class's line hole capital after selling all securities state is judged based on the recognition result.This method can integrate it is artificial report and the characteristics of video monitoring, realize the quick corresponding of regular bus and class line, and the accurately hole capital after selling all securities status information of acquisition class line, simultaneously, the class's of avoiding line adjusts the maintenance brought, classifies to the hole capital after selling all securities state of class's line, the programme planning for class's line provides reference frame.
Description
Technical field
The present invention relates to class's line hole capital after selling all securities condition adjudgements.More particularly, to a kind of class's line hole capital after selling all securities shape based on image recognition
State judgment method.
Background technology
In Express Logistics field, when regular bus in allocating after installing, the hole capital after selling all securities state for obtaining each regular bus is needed, and by class
Class's line that hole capital after selling all securities state and its of vehicle are run is mapped, and is the programme planning of class's line with the hole capital after selling all securities state of the class's of acquisition line, such as
Stop with extra bus etc., reference frame is provided.
Currently, for class's line hole capital after selling all securities state acquisition generally by manually reporting or video monitoring both of which, wherein：
One, artificial report refers to after regular bus installs, being estimated by the hole capital after selling all securities number that the operating personnel at scene are regular bus,
Then the hole capital after selling all securities number is corresponded on class line and is recorded and reported.Although artificial report can be good at corresponding with class's line of regular bus
Come, but due to the estimated value that hole capital after selling all securities number is operating personnel, the hole capital after selling all securities state accuracy being achieved in that is not high.Meanwhile to hole capital after selling all securities
The estimation of state varies with each individual, and lacks unified standard, is unfavorable for fining and digitized management.Further, on artificial
Class due to the case where lacking effectively supervision, hole capital after selling all securities state is disorderly reported be difficult to find, therefore bring verification etc. it is follow-up difficult.
Two, video monitoring refers to obtaining the corresponding surveillance video for allocating buckle by fixed operation personnel, by looking into
See surveillance video come the hole capital after selling all securities number of the class's of acquisition line.Video monitoring need operation personnel monitor whithin a period of time regular bus from
The moment opened, in this way and its labor intensive.Meanwhile actual production may adjust the correspondence of each bayonet and class's line；One
Bayonet may also correspond to multiple classes of lines.Since video monitoring can only check the video recording of fixed card buckle position, in regular bus and class
Line carry out to it is corresponding when need constantly to be adjusted according to the maintenance of buckle and class line, i.e., if without safeguarding in time, just can not find
Want the class's line looked for.
Accordingly, it is desirable to provide a kind of class's line hole capital after selling all securities state judging method based on image recognition.
Invention content
The purpose of the present invention is to provide the judgment method of a kind of class of line hole capital after selling all securities state, this method, which can integrate, manually to be reported
The characteristics of with video monitoring, realizes the quick corresponding of regular bus and class line, and the accurately hole capital after selling all securities status information of acquisition class line, meanwhile,
The class's of avoiding line adjusts the maintenance brought, classifies to the hole capital after selling all securities state of class's line, the programme planning for class's line provides reference frame.
In order to achieve the above objectives, the present invention uses following technical proposals：
A kind of class's line hole capital after selling all securities state judging method based on image recognition, this method include：
Step S1：Regular bus loading condition take pictures and obtains regular bus photo；
Step S2：The corresponding job order of regular bus is scanned and obtains regular bus information；
Step S3：Class's line information that regular bus corresponds to class's line is extracted in regular bus information, by regular bus photo and class's line information phase
Association；
Step S4：Regular bus photo is identified based on deep neural network, obtains recognition result；
Step S5：Associated class's line hole capital after selling all securities state is judged based on recognition result.
In the present invention, by the regular bus photo of the acquisition that regular bus loading condition will be carried out taking pictures with to the corresponding task of regular bus
The regular bus information for being singly scanned acquisition is associated, and it is corresponding with accurately facilitating for class line to realize regular bus, meanwhile, the regular bus of acquisition
Photo can avoid the wrong report problem during manually reporting, and improve the accuracy of hole capital after selling all securities state.Further, it is based on depth
Regular bus photo is identified in neural network, it can be ensured that there is unified standard to the judgement of class's line hole capital after selling all securities state, meanwhile, lead to
Cross the adjustment that could be adjusted to realize hole capital after selling all securities state classification to neural network.
Preferably, after regular bus installs, to regular bus loading condition take pictures with target rifle obtains regular bus photo.
Preferably, the corresponding job order of regular bus is scanned with target rifle and obtains regular bus information.
In the present invention, is robbed by target to be taken pictures to regular bus loading condition and be scanned to job order, can be facilitated,
Fast, accurately that regular bus photo is corresponding with class's line, reduce intermediate link, the effect that the class's of improving line hole capital after selling all securities state obtains
Rate.
Preferably, recognition result includes " expiring vehicle completely ", " empty 5m3Within ", " empty 5-30m3", " empty 30m3More than " and " nothing
Method identifies ".
In the application, recognition result is classified according to grade, conveniently intuitively class's line hole capital after selling all securities state can be carried out
The case where judging, while increasing " None- identified ", can adapt to the various situations in practical operation.
Preferably, associated class's line hole capital after selling all securities state is judged based on recognition result in step S5, including：
When recognition result is " expiring vehicle completely ", judge corresponding hole capital after selling all securities state for 0；
When recognition result is " empty 5m3Within " when, judge corresponding hole capital after selling all securities state for 2.5m3；
When recognition result is " empty 5-30m3" when, judge corresponding hole capital after selling all securities state for 17.5m3；
When recognition result is " empty 30m3 or more ", judge corresponding hole capital after selling all securities state for (30+V)/2m3；And
When recognition result is " None- identified ", corresponding hole capital after selling all securities state can not be judged, and record recognition result；
Wherein, V is the maximum volume of regular bus.
In the present invention, class's line hole capital after selling all securities state is judged corresponding to recognition result, obtains class's line hole capital after selling all securities of different stage
State, the hole capital after selling all securities status values of each rank represent the average value of the rank, judge gained hole capital after selling all securities state be directly used in as
The programme planning of class's line provides reference frame.
Preferably, further include after executing the step S1：
Judge whether regular bus photo is " photo lack of standardization ", if so, can not judge corresponding hole capital after selling all securities state, and the class of record
Licence piece.
It is further preferred that when licence piece on duty meets at least one following situations, judge regular bus photo for " photograph lack of standardization
Piece "：
The area shared by vehicle frame in regular bus photo accounts for the 25% or less of regular bus photo area；
The area shared by the vehicle frame of left and right in regular bus photo accounts for the 70% or less of left and right vehicle frame area；
The edge of frame of getting on or off the bus in regular bus photo is not included in regular bus photo；
The closing of the door of regular bus in regular bus photo；And
Short side can not be judged in regular bus photo.
In the present invention, specification situation judgement is carried out by the regular bus photo to acquisition, can reject in time and not conform to specification
Photo eliminates and does not conform to specification photo inflow subsequent step, improves treatment effeciency, it is ensured that deep neural network is to regular bus photo
The reliability being identified.It is possible to further determine regular bus photo by adjusting the criterion of " photo lack of standardization "
Qualification rate.
Preferably, deep neural network uses Xception network structures.In the present invention, using the Xception nets of Google
Network structure.
When being preferably based on deep neural network regular bus photo being identified, using hyper parameter search technique and image
Random perturbation technology.
In the present invention, using grid search thoughts to optimizer, pond channel type, data prediction type and pre- instruction
Practice four kinds of parameters to scan for, most has parameter combination with selection determination.In the training process by input picture Random-Rotation,
Light enhances the modes such as (decrease), random cropping and increases to model interference at random.The model after training is set to have better generalization
Can, avoid over-fitting.
When being preferably based on deep neural network regular bus photo being identified, built-up pattern technology is also used, wherein：
1 layer of output recognition result of prediction；
Obtain the loading information of regular bus；
2 layers of prediction is collectively formed based on the recognition result for loading the 1 layer of output of information and prediction, with right
Associated class's line hole capital after selling all securities state is judged.
In the present invention, using the Xception network structures of Google, and hyper parameter search technique, image random perturbation are combined
Technology and built-up pattern technology make accuracy rate be promoted from only with the 70% of Xception network structures to 90%, and having reached makes
Use standard.
Description of the drawings
Specific embodiments of the present invention will be described in further detail below in conjunction with the accompanying drawings.
Fig. 1 shows class's line hole capital after selling all securities state judging method block diagram based on image recognition.
Fig. 2 shows image recognition circuit theory schematic diagrams in one embodiment of the invention.
Fig. 3 shows image recognition circuit theory schematic diagram in another embodiment of the present invention.
Specific implementation mode
In order to illustrate more clearly of the present invention, the present invention is done further with reference to preferred embodiments and drawings
It is bright.Similar component is indicated with identical reference numeral in attached drawing.It will be appreciated by those skilled in the art that institute is specific below
The content of description is illustrative and be not restrictive, and should not be limited the scope of the invention with this.
In order to solve manually to report and the drawbacks of video monitoring, the present invention discloses a kind of class's line hole capital after selling all securities based on image recognition
State judging method, this method include：
Step S1：Regular bus loading condition take pictures and obtains regular bus photo.
Step S2：The corresponding job order of regular bus is scanned and obtains regular bus information；
Step S3：Class's line information that regular bus corresponds to class's line is extracted in regular bus information, by regular bus photo and class's line information phase
Association；
Step S4：Regular bus photo is identified based on deep neural network, obtains recognition result；
Step S5：Associated class's line hole capital after selling all securities state is judged based on recognition result.
In the present invention, by the regular bus photo of the acquisition that regular bus loading condition will be carried out taking pictures with to the corresponding task of regular bus
The regular bus information for being singly scanned acquisition is associated, and it is corresponding with accurately facilitating for class line to realize regular bus, meanwhile, the regular bus of acquisition
Photo can avoid the wrong report problem during manually reporting, and improve the accuracy of hole capital after selling all securities state.Further, it is based on depth
Regular bus photo is identified in neural network, it can be ensured that there is unified standard to the judgement of class's line hole capital after selling all securities state, meanwhile, lead to
Cross the adjustment that could be adjusted to realize hole capital after selling all securities state classification to neural network.
With reference to specific embodiment, class's line hole capital after selling all securities state judging method based on image recognition of the present invention is said
It is bright.
First, carry out execute-in-place.
First, when regular bus in allocating after installing, to regular bus loading condition take pictures with target rifle obtains regular bus photo, class
The actual load situation of regular bus is had recorded in licence piece.
In practice, specification is made to regular bus photo, if there are following five kinds of situations, assert that photo is photograph lack of standardization
Piece：
(1) area shared by the vehicle frame in regular bus photo accounts for the 25% or less of regular bus photo area；
(2) area shared by the left and right vehicle frame in regular bus photo accounts for the 70% or less of left and right vehicle frame area；
(3) edge of the frame of getting on or off the bus in regular bus photo is not included in regular bus photo；
(4) in regular bus photo regular bus closing of the door；
(5) it can not judge short side in regular bus photo.
Then, after obtaining regular bus photo, the corresponding job order of regular bus is scanned with target rifle and obtains regular bus information, it will
Regular bus photo is mapped with class line.It is robbed by target to be taken pictures and be scanned to job order, energy to regular bus loading condition
It is enough convenient, it is fast, accurately that regular bus photo is corresponding with class's line, reduce intermediate link, the class's of improving line hole capital after selling all securities state obtains
The efficiency taken.
Second, carry out Model Identification.
First, regular bus photo is identified based on deep neural network.Specifically, the photograph for needing to identify is transferred from system
Piece identifies that every photo is " expiring vehicle completely ", " empty 5m with deep neural network3Within ", " 5~30m of sky3", " empty 30m3More than "
Which kind of in " can not judge ".
In one embodiment, deep neural network uses the Xception network structures of Google, it should be noted that this hair
The bright type for being not intended to limit deep neural network, other network structures that can complete deep neural network identification function equally belong to
In the inventive concept of the present invention.
Recognition result is classified according to grade, conveniently intuitively class's line hole capital after selling all securities state can be judged, simultaneously
The case where increasing " None- identified " can adapt to the various situations in practical operation.
When regular bus photo being identified based on deep neural network, using hyper parameter search technique, image random perturbation
Technology and built-up pattern technology, wherein：
(1) hyper parameter search technique
It will be appreciated by persons skilled in the art that same network structure, can be arranged different parameters.Different parameters group
It is very big to close the performance difference on final result.
In the embodiment of the present invention, parameter in following four is scanned for using grid search thoughts, it is optimal to select
Parameter combination.
A) optimizers (optimizer)
B) ponds channel type (pooling)
C) data predictions mode (data preprocessing)
D) pre-training (pre-train&fine-tune)
(2) image random perturbation technology
In the embodiment of the present invention, in the training process by input picture Random-Rotation, light are enhanced at random (decrease),
The modes such as random cropping, which increase, interferes model.So that the model after training is had better Generalization Capability, avoids over-fitting.
(3) built-up pattern technology
In one embodiment, the frame of traditional problem of image recognition, nethermost Input are as shown in Figure 2
Regular bus image is finally predicted as " expiring vehicle completely ", " empty 5m by hidden layer3Within ", " 5~30m of sky3", " empty 30m3More than " and
Which kind of in " can not judge ".
For Express Logistics, other than regular bus photo, some other information can be obtained and carry out auxiliary judgment, such as：Regular bus fills
Holder amount, weight, collection packet number etc. can describe the information of regular bus loading in regular bus, therefore in another embodiment by model tune
It is whole as shown in Figure 3.
In figure 3, output " expiring vehicle completely ", " empty 5m at prediction1 layers3Within ", " 5~30m of sky3", " empty 30m3
More than " and " can not judge " in per a kind of probability.The information that can be described regular bus and load is added in input 2 (i.e. above
Regular bus carrier member amount, weight, collection packet number etc. in regular bus), be connected for 1 layer with prediction, place into together next
DNN (deep-neural-network) finally predicts " expiring vehicle completely ", " empty 5m again3Within ", " 5~30m of sky3", " empty 30m3More than " and
Which kind of in " can not judge " --- i.e. 2 layers of prediction.
Then, associated class's line hole capital after selling all securities state is judged based on recognition result, including：
When recognition result is " expiring vehicle completely ", judge corresponding hole capital after selling all securities state for 0；
When recognition result is " empty 5m3Within " when, judge corresponding hole capital after selling all securities state for 2.5m3；
When recognition result is " empty 5-30m3" when, judge corresponding hole capital after selling all securities state for 17.5m3；
When recognition result is " empty 30m3More than " when, judge corresponding hole capital after selling all securities state for (30+V)/2m3；And
When recognition result is " None- identified ", recognition result is pushed into operation personnel, personnel specification of taking pictures is urged to be taken pictures
And the personnel of taking pictures are examined for foundation with the recognition result of " None- identified "；
Wherein, V is the maximum volume of regular bus.
Class's line hole capital after selling all securities state is judged corresponding to recognition result, obtains class's line hole capital after selling all securities state of different stage, each
The hole capital after selling all securities status values of rank represent the average value of the rank, judge that the hole capital after selling all securities state of gained is directly used in as the scheme of class's line
Planning provides reference frame.Particularly, when it is " None- identified " that operating personnel's unrest bat or wrong bat, which cause recognition result, the same day is right
The regular bus loading answered will be unable to learn from image-recognizing method, be at this time according to taking pictures with the recognition result of " None- identified "
Personnel, which examine, can reduce disorderly bat and the wrong generation for clapping phenomenon.
In yet another embodiment, further include after executing the step S1：Judge whether regular bus photo is " photo lack of standardization ",
If so, can not judge corresponding hole capital after selling all securities state, regular bus photo is pushed into operation personnel, urge personnel specification of taking pictures take pictures and with
The regular bus photo of " photo lack of standardization " is that foundation examines the personnel of taking pictures.
Specification situation judgement is carried out by the regular bus photo to acquisition, the photo for not conforming to specification can be rejected in time, saved
Do not conform to specification photo and flow into subsequent step, improves treatment effeciency, it is ensured that regular bus photo is identified in deep neural network
Reliability.It is possible to further determine the qualification rate of regular bus photo by adjusting the criterion of " photo lack of standardization ".
In the embodiment of the present invention, is taken pictures with target rifle and the action for sweeping job order is convenient, it is quick and not error-prone, simultaneously
The correspondence of regular bus photo and class's line is also more accurate.By the hole capital after selling all securities of Model Identification photo, to providing the whole network unified standard
Standard, eliminate artificial subjective impact, this method be superior in efficiency, accuracy rate before method.
Each embodiment is described by the way of progressive in specification, the highlights of each of the examples are with other realities
Apply the difference of example, just to refer each other for identical similar portion between each embodiment.For device disclosed in embodiment
Speech, since it is corresponded to the methods disclosed in the examples, so description is fairly simple, related place is referring to method part illustration
?.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, depends on the specific application and design constraint of technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
Principle and implementation of the present invention are described for specific case used herein, and above example is said
The bright method and its core concept for being merely used to help understand the present invention.It should be pointed out that for the ordinary skill of the art
, without departing from the principle of the present invention, can be with several improvements and modifications are made to the present invention for personnel, these improvement
It is also fallen within the protection scope of the claims of the present invention with modification.
It should also be noted that, in the present specification, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only that
A little elements, but also include the other elements being not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or equipment including element.
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention may be used also on the basis of the above description for those of ordinary skill in the art
To make other variations or changes in different ways, all embodiments can not be exhaustive here, it is every to belong to this hair
Row of the obvious changes or variations that bright technical solution is extended out still in protection scope of the present invention.
Claims (10)
1. a kind of class's line hole capital after selling all securities state judging method based on image recognition, which is characterized in that the method includes：
Step S1：Regular bus loading condition take pictures and obtains the regular bus photo；
Step S2：The corresponding job order of the regular bus is scanned and obtains the regular bus information；
Step S3：Class's line information that the regular bus corresponds to class's line is extracted in the regular bus information, by the regular bus photo and institute
The class's of stating line information is associated；
Step S4：The regular bus photo is identified based on deep neural network, obtains recognition result；
Step S5：Associated class's line hole capital after selling all securities state is judged based on the recognition result.
2. according to claim 1 class of line hole capital after selling all securities state judging method, which is characterized in that after regular bus installs, with target rifle
Regular bus loading condition take pictures and obtains the regular bus photo.
3. according to claim 2 class of line hole capital after selling all securities state judging method, which is characterized in that with the target rifle to the regular bus
Corresponding job order, which is scanned, obtains the regular bus information.
4. according to claim 1 class of line hole capital after selling all securities state judging method, which is characterized in that the recognition result includes " complete
Full up vehicle ", " empty 5m3Within ", " empty 5-30m3", " empty 30m3More than " and " None- identified ".
5. class's line hole capital after selling all securities state judging method according to claim 4, which is characterized in that based on described in the step S5
Recognition result judges associated class's line hole capital after selling all securities state, including：
When recognition result is " expiring vehicle completely ", judge corresponding hole capital after selling all securities state for 0；
When recognition result is " empty 5m3Within " when, judge corresponding hole capital after selling all securities state for 2.5m3；
When recognition result is " empty 5-30m3" when, judge corresponding hole capital after selling all securities state for 17.5m3；
When recognition result is " empty 30m3More than " when, judge corresponding hole capital after selling all securities state for (30+V)/2m3；And
When recognition result is " None- identified ", corresponding hole capital after selling all securities state can not be judged, and record the recognition result；
Wherein, V is the maximum volume of the regular bus.
6. according to claim 1 class of line hole capital after selling all securities state judging method, which is characterized in that gone back after having executed the step S1
Including：
Judge whether the regular bus photo is " photo lack of standardization ", if so, can not judge corresponding hole capital after selling all securities state, and records institute
State regular bus photo.
7. the class's line hole capital after selling all securities state judging method stated according to claim 6, which is characterized in that when the regular bus photo meet it is following
When at least one situation, judge the regular bus photo for " photo lack of standardization "：
The area shared by vehicle frame in the regular bus photo accounts for the 25% or less of the regular bus photo area；
The area shared by the vehicle frame of left and right in the regular bus photo accounts for the 70% or less of the left and right vehicle frame area；
The edge of frame of getting on or off the bus in the regular bus photo is not included in the regular bus photo；
The closing of the door of regular bus described in the regular bus photo；And
It can not judge short side in the regular bus photo.
8. according to any class's line hole capital after selling all securities state judging method in claim 1-7, which is characterized in that the depth nerve
Network uses Xception network structures.
9. according to claim 8 class of line hole capital after selling all securities state judging method, which is characterized in that described to be based on deep neural network
When the regular bus photo is identified, using hyper parameter search technique and image random perturbation technology.
10. according to claim 8 class of line hole capital after selling all securities state judging method, which is characterized in that described to be based on depth nerve net
When the regular bus photo is identified in network, built-up pattern technology is also used, wherein：
1 layer of output recognition result of prediction；
Obtain the loading information of the regular bus；
2 layers of prediction is collectively formed based on the recognition result for loading the 1 layer of output of information and the prediction,
To judge associated class's line hole capital after selling all securities state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810346729.2A CN108805007A (en) | 2018-04-18 | 2018-04-18 | Class's line hole capital after selling all securities state judging method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810346729.2A CN108805007A (en) | 2018-04-18 | 2018-04-18 | Class's line hole capital after selling all securities state judging method based on image recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108805007A true CN108805007A (en) | 2018-11-13 |
Family
ID=64094363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810346729.2A Pending CN108805007A (en) | 2018-04-18 | 2018-04-18 | Class's line hole capital after selling all securities state judging method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108805007A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104036262A (en) * | 2014-06-30 | 2014-09-10 | 南京富士通南大软件技术有限公司 | Method and system for screening and recognizing LPR license plate |
| CN104931985A (en) * | 2015-05-25 | 2015-09-23 | 北京物通时空网络科技开发有限公司 | Intelligent vehicle load state judging and distribution positioning system |
| CN107103654A (en) * | 2017-05-02 | 2017-08-29 | 郑州华瑞伟业电子科技有限公司 | A kind of Cargo Inspection safety detection controlling and management and method |
| CN107122747A (en) * | 2017-04-28 | 2017-09-01 | 北京理工大学 | A kind of railway carriage state non-contact detection device and method |
| CN107590741A (en) * | 2017-09-19 | 2018-01-16 | 广东工业大学 | A kind of method and system of predicted pictures popularity |
-
2018
- 2018-04-18 CN CN201810346729.2A patent/CN108805007A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104036262A (en) * | 2014-06-30 | 2014-09-10 | 南京富士通南大软件技术有限公司 | Method and system for screening and recognizing LPR license plate |
| CN104931985A (en) * | 2015-05-25 | 2015-09-23 | 北京物通时空网络科技开发有限公司 | Intelligent vehicle load state judging and distribution positioning system |
| CN107122747A (en) * | 2017-04-28 | 2017-09-01 | 北京理工大学 | A kind of railway carriage state non-contact detection device and method |
| CN107103654A (en) * | 2017-05-02 | 2017-08-29 | 郑州华瑞伟业电子科技有限公司 | A kind of Cargo Inspection safety detection controlling and management and method |
| CN107590741A (en) * | 2017-09-19 | 2018-01-16 | 广东工业大学 | A kind of method and system of predicted pictures popularity |
Non-Patent Citations (2)
| Title |
|---|
| 张文龙: "基于CNN的载货列车信息识别系统设计与实现", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
| 米雪玉 等: "《矿山爆炸物品仓储与配送安全管理系统研究》", 31 January 2015, 哈尔滨工程大学出版社 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108710892B (en) | Cooperative immune defense method for multiple anti-picture attacks | |
| CN110837963A (en) | Risk control platform construction method based on data, model and strategy | |
| CN110929918A (en) | 10kV feeder line fault prediction method based on CNN and LightGBM | |
| CN108564052A (en) | Multi-cam dynamic human face recognition system based on MTCNN and method | |
| CN110139067A (en) | A kind of wild animal monitoring data management information system | |
| CN109492620A (en) | Monitoring device and its control device, post monitoring method and readable storage medium storing program for executing | |
| CN109766992A (en) | Industry control abnormality detection and attack classification based on deep learning | |
| CN104050361A (en) | Intelligent analysis early warning method for dangerousness tendency of prison persons serving sentences | |
| CN108875834A (en) | Image clustering method, device, computer equipment and storage medium | |
| CN108647691A (en) | A kind of image classification method based on click feature prediction | |
| CN109063759A (en) | A kind of neural network structure searching method applied to the more attribute forecasts of picture | |
| CN109347853A (en) | The method for detecting abnormality towards Integrated Electronic System based on depth Packet analyzing | |
| CN110009224A (en) | Suspect's violation probability prediction technique, device, computer equipment and storage medium | |
| CN106095939A (en) | The acquisition methods of account authority and device | |
| CN114357243A (en) | Massive real-time video stream multistage analysis and monitoring system | |
| CN116109139A (en) | Wind control strategy generation method, decision method, server and storage medium | |
| CN108375729B (en) | Degenerate state appraisal procedure is clustered based on the motor SOM that Fisher differentiates | |
| CN107437083A (en) | A kind of video behavior recognition methods of adaptive pool | |
| CN111930526A (en) | Load prediction method, load prediction device, computer equipment and storage medium | |
| CN113434575B (en) | Data attribution processing method, device and storage medium based on data warehouse | |
| CN108805007A (en) | Class's line hole capital after selling all securities state judging method based on image recognition | |
| KR102525491B1 (en) | Method of providing structure damage detection report | |
| CN114338351B (en) | Network anomaly root cause determination method and device, computer equipment and storage medium | |
| CN115796826A (en) | Management method, system, device and storage medium for ship safety management and control | |
| CN110008942A (en) | Certificate verification method, device, server and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181113 |