CN109523553A - A kind of container unusual fluctuation monitoring method based on LSD straight-line detection partitioning algorithm - Google Patents
A kind of container unusual fluctuation monitoring method based on LSD straight-line detection partitioning algorithm Download PDFInfo
- Publication number
- CN109523553A CN109523553A CN201811347793.9A CN201811347793A CN109523553A CN 109523553 A CN109523553 A CN 109523553A CN 201811347793 A CN201811347793 A CN 201811347793A CN 109523553 A CN109523553 A CN 109523553A
- Authority
- CN
- China
- Prior art keywords
- lsd
- unusual fluctuation
- algorithm
- template
- container
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Abstract
The invention discloses a kind of container unusual fluctuation monitoring methods based on LSD straight-line detection partitioning algorithm, including step A: generating standard form, step B: statistics is compared in new video frame straight line element, as a result step C is disposed.The present invention extracts graph line technological means using LSD algorithm and obtains initial position, the state of container by the video acquired to camera, if unusual fluctuation occurs for container, by the comparison with initial position, the container that unusual fluctuation occurs then is found out, and export unusual fluctuation event.The present invention effectively reduces the influence of camera shake by establishing multiframe template；And standard form is further established, sensitizing range delimited, the report manner of unusual fluctuation is effectively improved, multithreading divides, and improves treatment effeciency.
Description
Technical field
The present invention relates to container monitoring technical field, in particular to a kind of collection based on LSD straight-line detection partitioning algorithm
Vanning unusual fluctuation monitoring method.
Background technique
Container is fixed by the fixed device such as cage knob and binding bar, and is carried out according to lower heavy light stacking principle
It puts, to enhance the stability of container.If the stability of container is relatively low, rolling period is elongated, resists beam wind athwart sea
Ability is on the weak side, once container moves, ship gravity center instability can then be such that the risk of navigation increases, accident easily occurs.
In the past, the mode that the installation camera on ship is generallyd use in industry carries out the unusual fluctuation monitoring of container, but this
Kind mode can only observe the biggish unusual fluctuation of amplitude, and if there is subtle unusual fluctuation, camera is difficult to observe, and also not can avoid
The generation of container unusual fluctuation.
In recent years, industry begin one's study container unusual fluctuation monitoring new technology, LSD straight line monitor partitioning algorithm be exactly wherein
It is a kind of.LSD (Line Segment Detector) algorithm is a kind of linear feature detection algorithm, can be obtained in a very short period of time
The testing result of subpixel accuracy, LSD algorithm are designed to voluntarily control all without parameter regulation on the digital image
The quantity of erroneous detection processed.The straight linear profile in part, i.e. line segmentation in the algorithm detectable image.LSD algorithm executes rate
Fastly, precision is high, is highly suitable for the apparent container monitoring technology of linear feature.
Summary of the invention
To achieve the above object, the invention discloses a kind of, and the container unusual fluctuation based on LSD straight-line detection partitioning algorithm is supervised
Survey method, comprising the following steps: step A: generating standard form, and step B: statistics is compared in new video frame straight line element,
Step C, is as a result disposed；
The step A includes:
Step A1, video real-time is synchronous: for the video of input, multithreading being taken to handle, a thread is responsible for
Video data is taken, and converts data to the available matrix data of algorithm, another thread is responsible for event handling, to guarantee to locate
Data in lineation journey are latest frame, avoid causing video real-time poor due to terminal capabilities difference；
Step A2, it establishes blank template: according to the pixel size of video frame, generating an equal amount of blank template, ruler
It is very little are as follows: width is the width of input picture matrix, and the height of a height of input picture matrix, wave band number is 1, monochrome band, image position
Depth is 2bit；
Step A3, it extracts each frame video frame and carries out LSD algorithm processing: the video frame that camera acquires is converted into LSD
The data structure of algorithm, is handled by LSD algorithm, acquires the linear feature of current video frame, and straight line element confidence level is joined
Number being filled in the template of A2 with value 1 greater than 2, so far generates first frame template；
Step A4, LSD algorithm handles next frame video frame: it is identical as step A3, by straight line element with 1 fill mould of pixel value
Plate；50 frame of LSD algorithm continuous processing: circulation step A4 fills 50 straight line elements, it is therefore an objective to reduce single altogether in a template
The error that identification error and camera shake generate；
Step A5, will be being generated in A4 as a result, as immunity region；Element extension is carried out as unit of pixel, is divided into 5 grades
Prewarning area, as sensitizing range, the two collectively forms original template, is named as Mask herein；
The step B includes:
Step B1, event-template is created: identical as A2；
Step B2, LSD algorithm handles the 51st frame: it is identical as step A3, by straight line element with 1 filling template of pixel value, this
Place is named as Frame51；
Step B3, it counting: Frame51 is compared with Mask, the element information for falling in immunity region is not involved in statistics,
The element for falling in sensitizing range is for statistical analysis, counts the element quantity of sensitizing range respectively；
The step C includes:
Step C1, result judges: carrying out grade judgement according to pre-set threshold value, the region of unusual fluctuation is obtained, if different
Dynamic region is greater than the set value, and exports unusual fluctuation event according to the Forecast Mode of setting, algorithm end cycle is reset, again from A2
Start, carries out subsequent cycle；If be not greater than the set value, repetitive cycling executes step B；
As a further improvement of the above technical scheme, the step A3 is further included steps of
A31:LSD scaling factor；
A32:LSD gradient and direction template；
A33: gradient puppet sequence；
A34: linearity region increases；
The calculating of A35:NFA；
A36: dot density in class；
A37: optimization rectangle.
As a further improvement of the above technical scheme, the step C1 is further included steps of
C11: sensitivity threshold setting；
C12: result Forecast Mode setting；Setting output result pattern.
The beneficial effects of the present invention are:
Graph line technological means is extracted using LSD algorithm, by the video acquired to camera, obtains the first of container
Beginning position, state, if container occur unusual fluctuation, by the comparison with initial position, then find out occur unusual fluctuation container,
And export unusual fluctuation event.The present invention effectively reduces the influence of camera shake by establishing multiframe template；And it further establishes
Standard form delimit sensitizing range, effectively improve the report manner of unusual fluctuation, and multithreading divides, and improve treatment effeciency.
Detailed description of the invention
It to describe the technical solutions in the embodiments of the present invention more clearly, below will be to needed in the embodiment
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for ability
For the those of ordinary skill of domain, without any creative labor, it can also be obtained according to these attached drawings others
Attached drawing.
Fig. 1 is the working principle of the invention flow chart；
Fig. 2 is the working principle flow chart of LSD straight-line detection partitioning algorithm of the present invention；
Fig. 3 is the first frame original graph in experimentation of the present invention；
Fig. 4 is collected template first frame image in experimentation of the present invention；
Fig. 5 is the original template that completion is handled in experimentation of the present invention；
Fig. 6 is video frame of the container without unusual fluctuation obtained after comparing with original template；
Fig. 7 is the video frame that the container obtained after comparing with original template has unusual fluctuation；
Fig. 8 is the original graph for having unusual fluctuation video frame.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.
As shown in Figs. 1-2, a kind of container unusual fluctuation monitoring side based on LSD straight-line detection partitioning algorithm that the present invention provides
Method, this method include generating standard form and unusual fluctuation statistics and result disposition three parts.
First part generates standard form:
The method of the step A includes:
Step A1, video real-time is synchronous: for the video of input, multithreading being taken to handle, a thread is responsible for
Video data is taken, and converts data to the available matrix data of algorithm, another thread is responsible for event handling, to guarantee to locate
Data in lineation journey are latest frame, avoid causing video real-time poor due to terminal capabilities difference；
Step A2, it establishes blank template: according to the pixel size of video frame, generating an equal amount of blank template, ruler
It is very little are as follows: width is the width of input picture matrix, and the height of a height of input picture matrix, wave band number is 1, monochrome band, image position
Depth is 2bit；
Step A3, it extracts each frame video frame and carries out LSD algorithm processing: the video frame that camera acquires is converted into LSD
The data structure of algorithm, is handled by LSD algorithm, acquires the linear feature of current video frame, and straight line element confidence level is joined
Number being filled in the template of A2 with value 1 greater than 2, so far generates first frame template；Wherein LSD algorithm is handled, and can be divided into 7 streams
Journey: scaling, gradient and direction calculating, the sequence of gradient puppet, the growth of straight line (rectangle) region, the calculating of NFA, put in class it is close
Degree, optimization rectangle.By this 7 step, it can be detected from picture and be partitioned into straight line.The each step of brief description:
A31 scaling: the purpose of testing mesoscale zoom factor s=0.8 is to eliminate sawtooth effect.Then Gauss is used
The mode of down-sampling operates input picture；
A32 gradient and direction calculating: what is used when calculating is the template of 2*2, guarantees phase when in order to opposite module
The independence of adjoint point directional spreding；
The sequence of A33 gradient puppet: gradient value is bigger, significant marginal point, therefore is more suitable for seed point.To gradient
It is the very high work of timeliness that value, which carries out sequence completely, therefore gradient value is simply divided into 1024 grades, this
1024 grades cover gradient by 0~255 variation range, consumption when this sequence is a thread.Seed point is from gradient
It is worth highest grade to start to search for, successively down, until all the points are labeled as USED；
A34 straight line (rectangle) region increases: being met by seed point search angle and state is the point (eight neighborhood) of USED
The region of formation is known as line-supportregion.Meet the deflection of whole region in neighborhood in angle tolerance range t
The point of region-ang is added to this region；
The calculating of A35NFA: NFA (Number ofFalseALarms) is less than to judge the candidate rectangle of some in image
In opposite module in same position rectangle the quantity of mark point probability, NFA is bigger, shows current rectangle and phase in opposite module
It is more similar with position, opposite, current rectangle is more likely to be " real target "；
Dot density in A36 class: the calculating put in class will solve the shape that exists for be mistakenly identified as straight line of the curve because of t
Condition；
A37 optimizes rectangle: Rectangle Improvement.If current rectangle is not able to satisfy NFA still, use
1. reducing dian, 2. short side reduces a line, and 3. long side reduces a line, and 4. long side reduces another row, 5. reduces point p=p/2, until
Meet NFA.
Step A4, LSD algorithm handles next frame video frame: it is identical as step A3, by straight line element with 1 fill mould of pixel value
Plate；50 frame of LSD algorithm continuous processing: circulation step A4 fills 50 straight line elements, it is therefore an objective to reduce single altogether in a template
The error that identification error and camera shake generate；
Step A5, will be being generated in A4 as a result, as immunity region；Element extension is carried out as unit of pixel, is divided into 5 grades
Prewarning area, as sensitizing range, the two collectively forms original template, is named as Mask herein.
Statistics is compared in second part, new video frame straight line element:
The specific method of the step B includes:
Step B1, event-template is created: identical as A2；
Step B2, LSD algorithm handles the 51st frame: it is identical as step A3, by straight line element with 1 filling template of pixel value, this
Place is named as Frame51；
Step B3, it counting: Frame51 is compared with Mask, the element information for falling in immunity region is not involved in statistics,
The element for falling in sensitizing range is for statistical analysis, counts the element quantity of sensitizing range respectively.
Part III, result disposition:
Step C1, result judges: C11: sensitivity threshold setting；C12: result Forecast Mode setting；Setting output result sample
Formula (text, picture, sound) carries out grade judgement according to pre-set threshold value, obtains the region of unusual fluctuation, if unusual fluctuation
Region is greater than the set value, and exports unusual fluctuation event, algorithm end cycle according to the Forecast Mode of setting, and resetting restarts from A2,
Carry out subsequent cycle；If be not greater than the set value, repetitive cycling executes step B (B1, B2, B3).
EXPERIMENTAL EXEMPLIFICATIONThe explanation:
This experiment acquires video flowing using linux system, Haikang prestige view IP Camera, is divided based on LSD straight-line detection
The container unusual fluctuation monitoring method of algorithm.It is experimentation figure below:
Fig. 3 is the original graph of the first frame in experimentation, and Fig. 4 is that LSD algorithm handles collected mould in experimentation
Plate first frame, Fig. 5 are to handle the original template completed by LSD algorithm in experimentation, and Fig. 6 is after comparing with original template
Video frame of the obtained container without unusual fluctuation, Fig. 7 are the video frames that the container obtained after comparing with original template has unusual fluctuation, figure
8 be the original graph for having unusual fluctuation video frame.
According to original template compared with the video frame of no unusual fluctuation, without what difference, illustrate container without unusual fluctuation；According to first
For beginning template compared with the video frame for having unusual fluctuation, the discovery upper left container of video frame has unusual fluctuation.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Within mind and principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.Separately
Outside, the technical solution between each embodiment can be combined with each other, can be real with those of ordinary skill in the art but must be
Based on existing；It will be understood that the combination of this technical solution not when conflicting or cannot achieve when occurs in the combination of technical solution
In the presence of, also not the present invention claims protection scope within.
Claims (3)
1. a kind of container unusual fluctuation monitoring method based on LSD straight-line detection partitioning algorithm, which is characterized in that including step A: raw
At standard form, step B: statistics is compared in new video frame straight line element, as a result step C is disposed；
The step A includes:
Step A1, video real-time is synchronous: for the video of input, multithreading being taken to handle, a thread is responsible for taking view
Frequency evidence, and the available matrix data of algorithm is converted data to, another thread is responsible for event handling, to guarantee to handle line
Data in journey are latest frame, avoid causing video real-time poor due to terminal capabilities difference；
Step A2, it establishes blank template: according to the pixel size of video frame, generating an equal amount of blank template, size are as follows:
Width is the width of input picture matrix, and the height of a height of input picture matrix, wave band number is 1, monochrome band, and image locating depth is
2bit；
Step A3, it extracts each frame video frame and carries out LSD algorithm processing: the video frame that camera acquires is converted into LSD algorithm
Data structure, handled by LSD algorithm, acquire the linear feature of current video frame, and straight line element confidence level parameter is big
In 2 template for being filled in A2 with value 1, first frame template is so far generated；
Step A4, LSD algorithm handles next frame video frame: it is identical as step A3, by straight line element with 1 filling template of pixel value；
50 frame of LSD algorithm continuous processing: circulation step A4 fills 50 straight line elements altogether in a template, it is therefore an objective to reduce single identification
The error that error and camera shake generate；
Step A5, will be being generated in A4 as a result, as immunity region；Element extension is carried out as unit of pixel, is divided into 5 grades of early warning
Region, as sensitizing range, the two collectively forms original template, is named as Mask herein；
The step B includes:
Step B1, event-template is created: identical as A2；
Step B2, LSD algorithm handles the 51st frame: it is identical as step A3, by straight line element with 1 filling template of pixel value, order herein
Entitled Frame51；
Step B3, it counts: Frame51 being compared with Mask, the element information for falling in immunity region is not involved in statistics, falls in
The element of sensitizing range is for statistical analysis, counts the element quantity of sensitizing range respectively；
The step C includes:
Step C1, result judges: grade judgement is carried out according to pre-set threshold value, obtains the region of unusual fluctuation, if unusual fluctuation
Region is greater than the set value, and exports unusual fluctuation event, algorithm end cycle according to the Forecast Mode of setting, and resetting restarts from A2,
Carry out subsequent cycle；If be not greater than the set value, repetitive cycling executes step B.
2. the container unusual fluctuation monitoring method according to claim 1 based on LSD straight-line detection partitioning algorithm, feature exist
In the step A3 is further included steps of
A31:LSD scaling factor；
A32:LSD gradient and direction template；
A33: gradient puppet sequence；
A34: linearity region increases；
The calculating of A35:NFA；
A36: dot density in class；
A37: optimization rectangle.
3. the container unusual fluctuation monitoring method according to claim 1 based on LSD straight-line detection partitioning algorithm, feature exist
In the step C1 is further included steps of
C11: sensitivity threshold setting；
C12: result Forecast Mode setting；Setting output result pattern.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811347793.9A CN109523553B (en) | 2018-11-13 | 2018-11-13 | Container abnormal movement monitoring method based on LSD linear detection segmentation algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811347793.9A CN109523553B (en) | 2018-11-13 | 2018-11-13 | Container abnormal movement monitoring method based on LSD linear detection segmentation algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109523553A true CN109523553A (en) | 2019-03-26 |
| CN109523553B CN109523553B (en) | 2022-10-18 |
Family
ID=65776173
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811347793.9A Active CN109523553B (en) | 2018-11-13 | 2018-11-13 | Container abnormal movement monitoring method based on LSD linear detection segmentation algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109523553B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110246123A (en) * | 2019-05-21 | 2019-09-17 | 贵州民族大学 | A kind of tile laying regularity detection method based on machine vision |
| CN112053360A (en) * | 2020-10-10 | 2020-12-08 | 腾讯科技（深圳）有限公司 | Image segmentation method and device, computer equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090238486A1 (en) * | 2006-10-18 | 2009-09-24 | Paul Biezen | Image processing system, method and computer program for contrast enhancement of images |
| CN102833511A (en) * | 2012-09-18 | 2012-12-19 | 浙江红苹果电子有限公司 | Method and device for storing high-definition videos based on serial digital interface (SDI) |
| CN106599885A (en) * | 2016-08-30 | 2017-04-26 | 中海网络科技股份有限公司 | Monitoring system and method for container Bay |
| CN106986272A (en) * | 2017-02-24 | 2017-07-28 | 北京航天自动控制研究所 | It is a kind of to prevent slinging method and system based on the container container car that machine vision is tracked |
| CN107527347A (en) * | 2017-10-11 | 2017-12-29 | 南京大学 | Harbour container based on computer visual image processing lifts by crane safety monitoring method |
| CN107734268A (en) * | 2017-09-18 | 2018-02-23 | 北京航空航天大学 | A kind of structure-preserved wide baseline video joining method |
-
2018
- 2018-11-13 CN CN201811347793.9A patent/CN109523553B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090238486A1 (en) * | 2006-10-18 | 2009-09-24 | Paul Biezen | Image processing system, method and computer program for contrast enhancement of images |
| CN102833511A (en) * | 2012-09-18 | 2012-12-19 | 浙江红苹果电子有限公司 | Method and device for storing high-definition videos based on serial digital interface (SDI) |
| CN106599885A (en) * | 2016-08-30 | 2017-04-26 | 中海网络科技股份有限公司 | Monitoring system and method for container Bay |
| CN106986272A (en) * | 2017-02-24 | 2017-07-28 | 北京航天自动控制研究所 | It is a kind of to prevent slinging method and system based on the container container car that machine vision is tracked |
| CN107734268A (en) * | 2017-09-18 | 2018-02-23 | 北京航空航天大学 | A kind of structure-preserved wide baseline video joining method |
| CN107527347A (en) * | 2017-10-11 | 2017-12-29 | 南京大学 | Harbour container based on computer visual image processing lifts by crane safety monitoring method |
Non-Patent Citations (1)
| Title |
|---|
| 李涛等: "基于改进线段分割检测的电线杆遮挡检测算法", 《计算机工程》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110246123A (en) * | 2019-05-21 | 2019-09-17 | 贵州民族大学 | A kind of tile laying regularity detection method based on machine vision |
| CN110246123B (en) * | 2019-05-21 | 2021-08-03 | 贵州民族大学 | Tile paving regularity detection method based on machine vision |
| CN112053360A (en) * | 2020-10-10 | 2020-12-08 | 腾讯科技（深圳）有限公司 | Image segmentation method and device, computer equipment and storage medium |
| CN112053360B (en) * | 2020-10-10 | 2023-07-25 | 腾讯科技（深圳）有限公司 | Image segmentation method, device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109523553B (en) | 2022-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110363158B (en) | Millimeter wave radar and visual cooperative target detection and identification method based on neural network | |
| CN112766274B (en) | Water gauge image water level automatic reading method and system based on Mask RCNN algorithm | |
| EP3680609A1 (en) | Antenna downward inclination angle measurement method based on multi-scale deep semantic segmentation network | |
| CN107490398B (en) | A kind of gauge pointer automatic identifying method | |
| CN110084165A (en) | The intelligent recognition and method for early warning of anomalous event under the open scene of power domain based on edge calculations | |
| CN103996049B (en) | Ship overlength and overwidth detection method based on video image | |
| CN112798811B (en) | Speed measurement method, device and equipment | |
| CN110516636A (en) | A kind of monitoring method of process, device, computer equipment and storage medium | |
| CN111275960A (en) | Traffic road condition analysis method, system and camera | |
| CN106778570B (en) | A kind of pedestrian detection and tracking in real time | |
| CN102768022A (en) | Tunnel surrounding rock deformation detection method adopting digital camera technique | |
| CN109523553A (en) | A kind of container unusual fluctuation monitoring method based on LSD straight-line detection partitioning algorithm | |
| CN107067755A (en) | A kind of method for calibrating traffic monitoring camera automatically based on computer vision | |
| CN109982044B (en) | Tracking method of target positioning and tracking system based on CCTV sensor network | |
| CN110009675A (en) | Generate method, apparatus, medium and the equipment of disparity map | |
| CN106504227B (en) | Demographic method and its system based on depth image | |
| CN109360225A (en) | A kind of optimization system and method for motion model | |
| CN110070106A (en) | Smog detection method, device and electronic equipment | |
| CN115588040A (en) | System and method for counting and positioning coordinates based on full-view imaging points | |
| CN116844147A (en) | Pointer instrument identification and abnormal alarm method based on deep learning | |
| Du et al. | An open data platform for traffic parameters measurement via multirotor unmanned aerial vehicles video | |
| CN109165553A (en) | Petroleum installations remote sensing automatic identifying method and device based on deep learning | |
| CN109492573A (en) | A kind of pointer read method and device | |
| CN105335723B (en) | A kind of model recognizing method based on infrared structure light | |
| Reiterer et al. | A 3D optical deformation measurement system supported by knowledge-based and learning techniques |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |