CN108960052A - Ship overload detecting method based on video flowing - Google Patents
Ship overload detecting method based on video flowing Download PDFInfo
- Publication number
- CN108960052A CN108960052A CN201810521329.0A CN201810521329A CN108960052A CN 108960052 A CN108960052 A CN 108960052A CN 201810521329 A CN201810521329 A CN 201810521329A CN 108960052 A CN108960052 A CN 108960052A
- Authority
- CN
- China
- Prior art keywords
- target
- ship
- passenger
- target ship
- motion
- 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
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
A kind of ship overload detecting method based on video flowing includes the following steps: S1, carries out moving object detection to imaging area, chooses target ship；S2, the FAST set of characteristic points for extracting target ship, obtain target tracking result；S3, the traffic direction that target ship is judged according to target tracking result；S4, the traffic direction progress corresponding operating according to target ship, S5, judge that semaphore shipIsNull=True's is true and false, and carry out corresponding operating according to judging result；S6, start ship time up and down when target ship reaches harbour, passenger, update the number on target ship；S7, target ship restarting before, check object ship only whether overload.The present invention is based on target identifications and tracing algorithm, and the two field pictures for being separated by a period of time are chosen in live video stream, pre-processed to this two field pictures, extracted feature and identify, provide judgment basis for ship overload detection.
Description
Technical field
The present invention relates to a kind of detection method more particularly to a kind of ship overload detecting methods based on video flowing, belong to
Target identification and tracer technique field.
Background technique
Intelligent video monitoring is that computer vision field comes in develop an application direction very fast, that research is more.Its energy
Enough collected vision signal is handled, analyzed and understood using computer vision technique, and based on this to video
Monitoring system is controlled, so that video monitoring system be made to have preferably intelligence and robustness.In current intelligent video
In monitoring field, target tracking technology is one of most common technological means.So-called target tracking is exactly in the every of video image
The position for the moving target that user needs is determined in one frame image, and corresponds to same target in the image of different frame
Come.
Currently, the problem of China's ship carrying overloads, and especially the ship carrying of rural area ferry area overloads increasingly shows, benefit
Complete the detection that overloads to ship with target tracking technology, also gradually become target tracking technology main application direction it
One.But in the actual application of target tracking technology, because of the influence of the factors such as environment, light, more be stranded can be faced
Difficulty, such as appearance deformation, quick movement, the similar interference of background, plane internal rotation, plane external rotation, are blocked and are gone out at illumination variation
Visual field etc..When especially by the technical application in the overweight detection process of ship, it may appear that the feelings that ship or passenger can be blocked
Condition, so as to cause that cannot accurately identify and track.
With the continuous development of technology, occurs a kind of target tracking algorism based on characteristic matching at present, use is this
Target tracking algorism based on characteristic matching, even if there is the case where partial occlusion in target in scene, if characteristic point as it can be seen that
Still it may be implemented to continue tracking.The target tracking method needs to extract clarification of objective, and finding in each frame image should
Feature；The process of searching is exactly the process of characteristic matching.Target tracking algorithm based on characteristic matching is needed using characteristics of image
Expressive Features of the point as moving target, the extraction algorithm of characteristic point generally comprise: Scale Invariant Feature
Transform(SIFT)、Speeded Up Robust Features(SURF)、Features From Accelerated
Segment Test (FAST), ORiented Brief (ORB) etc..
Wherein, SIFT unstable, the weakness such as feature deficiency that improve previous feature point extraction, and to rotation translation scaling
There is robustness.SURF algorithm can regard the important improvement of SIFT algorithm as, solve SIFT computation complexity, time-consuming
Long disadvantage；But robustness of the SURF algorithm in terms of illumination and deformation will be inferior to SIFT.FAST is most successful in recent years
One of feature point extraction algorithm, other algorithms such as ORB algorithm can be regarded as based on FAST algorithm development.
Exactly because how the also excellent practical characteristic of FAST algorithm, make full use of this algorithm, better solve mesh
It marks tracer technique and is applied to this industry problem in the overweight detection process of ship, also just become those skilled in the art needs jointly
One of solve the problems, such as.
Summary of the invention
In view of the prior art there are drawbacks described above, the ship overload inspection based on video flowing that the purpose of the present invention is to propose to a kind of
Survey method.
Specifically, including the following steps:
S1, moving object detection is carried out to imaging area, chooses target ship, initializing signal amount shipIsNull=
True；
S2, the FAST set of characteristic points for extracting target ship, are obtained using the target tracking algorism based on Feature Points Matching
Target tracking result；
S3, the traffic direction of target ship is judged according to target tracking result to pull in shore or offshore；
S4, the traffic direction progress corresponding operating according to target ship,
If the direction of motion of target ship is to pull in shore, enter S5 step,
If the direction of motion of target ship is offshore, other ships that pull in shore is waited to enter imaging area, and returns to S1 step
Suddenly, it re-starts moving object detection and target ship is selected；
S5, judge that semaphore shipIsNull=True's is true and false, and carry out corresponding operating according to judging result,
If semaphore shipIsNull=True is that very, the number on initialized target ship is 0, and setting signal amount
ShipIsNull=False,
If semaphore shipIsNull=True is vacation, the existing passenger people on target ship is directly obtained from server
Number；
S6, start ship time up and down when target ship reaches harbour, passenger, carry out moving object detection again, choose target
Passenger judges the direction of motion of target passenger, and the direction of motion according to target passenger, carries out phase to the number on target ship
It should increase and decrease, update the number on target ship；
S7, target ship restarting before, check object ship only whether overload,
If number is more than limit for tonnage number on target ship, give a warning, prompts stand by passenger to go ashore, and according to going ashore
Ridership updates the patronage on target ship,
If number is less than limit for tonnage number on target ship, the number on target ship at this time is returned into server.
Preferably, moving object detection is carried out to imaging area described in S1, specifically comprised the following steps:
S11, pretreatment carry out denoising and brightness normalized to each frame image；
S12, background modeling, using mean filter method auto-initiation background, and constantly real-time update background；
S13, target detection are obtained on the basis of background subtraction commercial affairs using the threshold segmentation method based on two-dimentional cross entropy
Moving target；
S14, post-processing, with smear and connected region differentiate processing removal moving object detection result in noise spot and
The hole region of target internal.
3, the ship overload detecting method according to claim 1 based on video flowing, which is characterized in that described in S2
The FAST set of characteristic points for extracting target ship, specifically comprise the following steps:
Pixel in S21, selected digital image, and using this pixel as the center of circle, select the round field mould that a radius is r
Plate x；
Whether the selected pixel of S22, judgement is FAST characteristic point, and judgment formula is,
Wherein, p is the pixel to be calculated, and x is round domain template, and l (x) and l (p) are respectively the pixel of x and p
Value, ε d are the threshold value of color difference,
When at least N0 point is continuously then to recognize without interruption in N number of point that N is greater than on threshold value N0 and round domain template
It is FAST characteristic point for p；
S23, S21 step, S22 step are repeated, until completing the extraction of FAST set of characteristic points.
Preferably, the r=3, N0=12.
Preferably, target passenger is chosen described in S6, judges the direction of motion of target passenger, and the fortune according to target passenger
Dynamic direction, accordingly increases and decreases the number on target ship, is included the following steps:
Target passenger is chosen, the characteristic point of target passenger is extracted, is obtained using the target tracking algorism based on Feature Points Matching
The direction of motion to target passenger goes on board or goes ashore,
If the direction of motion of target passenger is to go on board, the number on target ship is added one,
If the direction of motion of target passenger is to go ashore, the number on target ship is subtracted one.
Compared with prior art, advantages of the present invention is mainly reflected in the following aspects:
The present invention is based on target identifications and tracing algorithm, and the two frame figures for being separated by a period of time are chosen in live video stream
Picture pre-processes this two field pictures, extracts feature and identify, fully ensure that the accuracy of identification process, be
Ship overload detection provides accurate judgment basis.The present invention can judge whether ship is to use for the first time on the same day simultaneously, root
It is judged that the difference of result, the acquisition modes of number and the process of overload detection are also not quite similar on ship, ensure that this
Detection processing effect of the invention when facing all kinds of different situations, ensure that practicability of the invention.In addition, detection of the invention
Method can also carry out expansion extension other than the problem of being able to solve the overload of each department ferry ship on this basis, use
In other detections in field of traffic, there is very wide application prospect.
In conclusion the invention proposes a kind of ship overload detecting method based on video flowing, has very high use
And promotional value.
Just attached drawing in conjunction with the embodiments below, the embodiment of the present invention is described in further detail, so that of the invention
Technical solution is more readily understood, grasps.
Detailed description of the invention
Fig. 1 is flow chart of the present invention；
Fig. 2 is FAST characteristic point schematic diagram in the present invention.
Specific embodiment
As shown in Figure 1, present invention discloses a kind of ship overload detecting method based on video flowing.
Specifically, including the following steps:
S1, moving object detection is carried out to imaging area, chooses target ship, initializing signal amount shipIsNull=
True。
S2, the FAST set of characteristic points for extracting target ship, are obtained using the target tracking algorism based on Feature Points Matching
Target tracking result.
S3, the traffic direction of target ship is judged according to target tracking result to pull in shore or offshore.
S4, the traffic direction progress corresponding operating according to target ship,
If the direction of motion of target ship is to pull in shore, enter S5 step,
If the direction of motion of target ship is offshore, other ships that pull in shore is waited to enter imaging area, and returns to S1 step
Suddenly, it re-starts moving object detection and target ship is selected.
S5, judge that semaphore shipIsNull=True's is true and false, and carry out corresponding operating according to judging result,
If semaphore shipIsNull=True is that very, the number on initialized target ship is 0, and setting signal amount
ShipIsNull=False,
If semaphore shipIsNull=True is vacation, the existing passenger people on target ship is directly obtained from server
Number.
Judge only whether Kong necessity is that target ship evening before that day can select in pier to object ship herein,
It then is not no passenger on ship when ship begins to use within second day, therefore when ship is sky can initialize passenger on board and be
Zero.
S6, start ship time up and down when target ship reaches harbour, passenger, carry out moving object detection again, choose target
Passenger judges the direction of motion of target passenger, and the direction of motion according to target passenger, carries out phase to the number on target ship
It should increase and decrease, update the number on target ship.
S7, target ship restarting before, check object ship only whether overload,
If number is more than limit for tonnage number on target ship, give a warning, prompts stand by passenger to go ashore, and according to going ashore
Ridership updates the patronage on target ship,
If number is less than limit for tonnage number on target ship, the number on target ship at this time is returned into server.
Moving object detection is carried out to imaging area described in S1, i.e., moving target is detected using background subtraction,
Detection process mainly includes four pretreatment, background modeling, target detection and post-processing steps.It specifically includes:
S11, pretreatment carry out denoising and brightness normalized to each frame image, to inhibit illuminance abrupt variation and noise
Influence.
S12, background modeling, using improved mean filter method auto-initiation background, and constantly real-time update background, with
Overcoming influences caused by ambient lighting variation.
S13, target detection are obtained on the basis of background subtraction commercial affairs using the threshold segmentation method based on two-dimentional cross entropy
Moving target.
S14, post-processing, with smear and connected region differentiate processing removal moving object detection result in noise spot and
The hole region of target internal.
The FAST set of characteristic points that target ship is extracted described in S2, specifically comprise the following steps:
Pixel in S21, selected digital image, and using this pixel as the center of circle, select the round field mould that a radius is r
Plate x, in the present embodiment, the r=3.
Whether the selected pixel of S22, judgement is FAST characteristic point, and FAST characteristic point refers to adjacent around the pixel
In domain, there is the color difference of enough pixel and the point larger, such as A, B two o'clock in Fig. 2, judgment formula is,
Wherein, p is the pixel to be calculated, and x is round domain template, and l (x) and l (p) are respectively the pixel of x and p
Value, ε d are the threshold value of color difference,
When at least N0 point is continuously then to recognize without interruption in N number of point that N is greater than on threshold value N0 and round domain template
It is FAST characteristic point for p.
In general, as r=3, N0=12.In order to improve the detection efficiency of FAST characteristic point, attachment can also be used
Accelerated tecnicality, first determine whether condition is met by the pixel of 4 apexes that 90 ° divide on the circumference of candidate point, referring to
1,5,9,13 this 4 points in Fig. 2.If at least 3 points meet condition in this 4 points, i.e., differed with the pixel value of candidate point
Greater than the threshold value of setting, then continue to judge other points on circumference.Otherwise being considered as the candidate point not is FAST characteristic point.
S23, S21 step, S22 step are repeated, until completing the extraction of FAST set of characteristic points.
Target passenger is chosen described in S6, judges the direction of motion of target passenger, and the direction of motion according to target passenger,
Number on target ship is accordingly increased and decreased, is included the following steps:
Target passenger is chosen, the characteristic point of target passenger is extracted, is obtained using the target tracking algorism based on Feature Points Matching
The direction of motion to target passenger goes on board or goes ashore,
If the direction of motion of target passenger is to go on board, the number on target ship is added one,
If the direction of motion of target passenger is to go ashore, the number on target ship is subtracted one.
It should be added that, it is contemplated that real-time, selection FAST characteristic point are tracked to carry out ship tracking and pedestrian,
In the treatment process of algorithm, the region of search of target in present frame is determined according to the tracking result of previous frame, further according to above
Description method extract region of search FAST characteristic point, then by the FAST characteristic point in current search region and previous frame mesh
FAST characteristic point in mark is matched.
Matching process is as follows: calculating and wants each characteristic point in matched two FAST set of characteristic points special at another
The arest neighbors in point set and time neighbour are levied, if the distance between characteristic point and arest neighbors arrive between at a distance from time neighbour with it
When ratio is less than certain threshold value, it is considered as the initial matching that arest neighbors is sample.
After choosing the target to be tracked by moving object detection algorithm, the boundary rectangle frame of the target can be obtained
R1,
The FAST set of characteristic points O1 for adopting extraction target with the aforedescribed process completes the initial of algorithm by above-mentioned two step
Chemical industry is made.During being tracked to target, it is assumed that the video image currently obtained is Ii frame, is obtained in I i-1 frame image
The target boundary rectangle frame obtained is Ri-1, and the FAST feature point set of target is combined into Oi-1, since the time interval between two frames has
Limit, the moving range of target is also small distance, carries out the appropriate target search region amplified and can obtained in Ii to Ri-1,
FAST set of characteristic points O is extracted in region of search, O matched with Oi-1, obtain matching point set to Mi-1, Mi, then Ii and
The displacement of target can be estimated by following formula between Ii-1 frame:
Δ P=Pi-Pi-1,
Wherein, Pi, Pi-1 are the orthocenter of Mi and Mi-1 respectively.Target also will appear scale and rotation often during the motion
Transformationization,
Scale of the target between Ii-1 and Ii and rotationally-varying can be calculated according to following two formula:
Wherein, aiAnd ajIt is MiMiddle any two point, biAnd bjIt is Mi-1It is middle respectively with ai, ajCorresponding two points, s are to estimate
The target scale zoom factor counted out, θ are the rotation angle rotated from the (i-1)-th frame to i frame, and R is spin matrix, the meter of θ
Calculation method is as follows:
Estimate the displacement of target, scale and it is rotationally-varying after, so that it may calculate the target in Ii+1.
Vi=sRVi-1+ Δ P,
Vi-1 therein is the vertex Ri-1, and the boundary rectangle frame of target in Ii frame can be obtained according to Vi.When feature in Mi
When point number is less than certain threshold value, chooses the characteristic point being located in Ri in O and be added in Mi, obtain clarification of objective point in Ii frame
Set Oi.
The present invention is based on target identifications and tracing algorithm, and the two frame figures for being separated by a period of time are chosen in live video stream
Picture pre-processes this two field pictures, extracts feature and identify, fully ensure that the accuracy of identification process, be
Ship overload detection provides accurate judgment basis.The present invention can judge whether ship is to use for the first time on the same day simultaneously, root
It is judged that the difference of result, the acquisition modes of number and the process of overload detection are also not quite similar on ship, ensure that this
Detection processing effect of the invention when facing all kinds of different situations, ensure that practicability of the invention.In addition, detection of the invention
Method can also carry out expansion extension other than the problem of being able to solve the overload of each department ferry ship on this basis, use
In other detections in field of traffic, there is very wide application prospect.
In conclusion the invention proposes a kind of ship overload detecting method based on video flowing, has very high use
And promotional value.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case where without departing substantially from spirit and essential characteristics of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included within the present invention, and any reference signs in the claims should not be construed as limiting the involved claims.
In addition, it should be understood that although this specification is described in terms of embodiments, but not each embodiment is only wrapped
Containing an independent technical solution, this description of the specification is merely for the sake of clarity, and those skilled in the art should
It considers the specification as a whole, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art
The other embodiments being understood that.
Claims (5)
1. a kind of ship overload detecting method based on video flowing, which comprises the steps of:
S1, moving object detection is carried out to imaging area, chooses target ship, initializing signal amount shipIsNull=True；
S2, the FAST set of characteristic points for extracting target ship, obtain target using the target tracking algorism based on Feature Points Matching
Track result；
S3, the traffic direction of target ship is judged according to target tracking result to pull in shore or offshore；
S4, the traffic direction progress corresponding operating according to target ship,
If the direction of motion of target ship is to pull in shore, enter S5 step,
If the direction of motion of target ship is offshore, other ships that pull in shore are waited to enter imaging area, and return to S1 step, weight
New progress moving object detection and target ship are selected；
S5, judge that semaphore shipIsNull=True's is true and false, and carry out corresponding operating according to judging result,
If semaphore shipIsNull=True is that very, the number on initialized target ship is 0, and setting signal amount
ShipIsNull=False,
If semaphore shipIsNull=True is vacation, the existing patronage on target ship is directly obtained from server；
S6, start ship time up and down when target ship reaches harbour, passenger, carry out moving object detection again, choose target passenger,
Judge the direction of motion of target passenger, and the direction of motion according to target passenger, the number on target ship is accordingly increased
Subtract, updates the number on target ship；
S7, target ship restarting before, check object ship only whether overload,
If number is more than limit for tonnage number on target ship, give a warning, prompts stand by passenger to go ashore, and according to the passenger that goes ashore
Number updates the patronage on target ship,
If number is less than limit for tonnage number on target ship, the number on target ship at this time is returned into server.
2. the ship overload detecting method according to claim 1 based on video flowing, which is characterized in that taking the photograph described in S1
As region progress moving object detection, specifically comprise the following steps:
S11, pretreatment carry out denoising and brightness normalized to each frame image；
S12, background modeling, using mean filter method auto-initiation background, and constantly real-time update background；
S13, target detection obtain movement using the threshold segmentation method based on two-dimentional cross entropy on the basis of background subtraction commercial affairs
Target；
S14, post-processing differentiate that processing removes noise spot and target in moving object detection result with connected region with smearing
Internal hole region.
3. the ship overload detecting method according to claim 1 based on video flowing, which is characterized in that extracted described in S2
The FAST set of characteristic points of target ship, specifically comprise the following steps:
Pixel in S21, selected digital image, and using this pixel as the center of circle, select the round domain template x that a radius is r；
Whether the selected pixel of S22, judgement is FAST characteristic point, and judgment formula is,
Wherein, p is the pixel to be calculated, and x is round domain template, and l (x) and l (p) are respectively the pixel value of x and p, ε
D is the threshold value of color difference,
When at least N0 point is continuously without interruption in N number of point that N is greater than on threshold value N0 and round domain template, then it is assumed that p
For FAST characteristic point；
S23, S21 step, S22 step are repeated, until completing the extraction of FAST set of characteristic points.
4. the ship overload detecting method according to claim 3 based on video flowing, it is characterised in that: the r=3, N0
=12.
5. the ship overload detecting method according to claim 1 based on video flowing, which is characterized in that chosen described in S6
Target passenger judges the direction of motion of target passenger, and the direction of motion according to target passenger, to the number on target ship into
The corresponding increase and decrease of row, includes the following steps:
Target passenger is chosen, the characteristic point of target passenger is extracted, mesh is obtained using the target tracking algorism based on Feature Points Matching
The direction of motion of scalar multiplication visitor goes on board or goes ashore,
If the direction of motion of target passenger is to go on board, the number on target ship is added one,
If the direction of motion of target passenger is to go ashore, the number on target ship is subtracted one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810521329.0A CN108960052A (en) | 2018-05-28 | 2018-05-28 | Ship overload detecting method based on video flowing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810521329.0A CN108960052A (en) | 2018-05-28 | 2018-05-28 | Ship overload detecting method based on video flowing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108960052A true CN108960052A (en) | 2018-12-07 |
Family
ID=64492221
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810521329.0A Pending CN108960052A (en) | 2018-05-28 | 2018-05-28 | Ship overload detecting method based on video flowing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108960052A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110349389A (en) * | 2019-07-12 | 2019-10-18 | 武汉理工大学 | A kind of safety monitoring method of the small-sized passenger ferry in inland river |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103021059A (en) * | 2012-12-12 | 2013-04-03 | 天津大学 | Video-monitoring-based public transport passenger flow counting method |
| CN104442834A (en) * | 2013-09-12 | 2015-03-25 | 侯庆生 | Subway seating area number-of-passenger control device and method |
| CN104966045A (en) * | 2015-04-02 | 2015-10-07 | 北京天睿空间科技有限公司 | Video-based airplane entry-departure parking lot automatic detection method |
| CN106203276A (en) * | 2016-06-30 | 2016-12-07 | 中原智慧城市设计研究院有限公司 | A kind of video passenger flow statistical system and passenger flow statistical method |
| CN107657668A (en) * | 2017-09-22 | 2018-02-02 | 陕西师范大学 | Bus intelligent docking method |
-
2018
- 2018-05-28 CN CN201810521329.0A patent/CN108960052A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103021059A (en) * | 2012-12-12 | 2013-04-03 | 天津大学 | Video-monitoring-based public transport passenger flow counting method |
| CN104442834A (en) * | 2013-09-12 | 2015-03-25 | 侯庆生 | Subway seating area number-of-passenger control device and method |
| CN104966045A (en) * | 2015-04-02 | 2015-10-07 | 北京天睿空间科技有限公司 | Video-based airplane entry-departure parking lot automatic detection method |
| CN106203276A (en) * | 2016-06-30 | 2016-12-07 | 中原智慧城市设计研究院有限公司 | A kind of video passenger flow statistical system and passenger flow statistical method |
| CN107657668A (en) * | 2017-09-22 | 2018-02-02 | 陕西师范大学 | Bus intelligent docking method |
Non-Patent Citations (1)
| Title |
|---|
| 刘文萍 等: "一种新的背景减运动目标检测方法", 《计算机工程与应用》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110349389A (en) * | 2019-07-12 | 2019-10-18 | 武汉理工大学 | A kind of safety monitoring method of the small-sized passenger ferry in inland river |
| CN110349389B (en) * | 2019-07-12 | 2021-03-30 | 武汉理工大学 | Safety monitoring method for small inland-river passenger ferry |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106683119B (en) | Moving vehicle detection method based on aerial video image | |
| Bernecker et al. | Continuous short-term irradiance forecasts using sky images | |
| Benedek et al. | Bayesian foreground and shadow detection in uncertain frame rate surveillance videos | |
| US20120328161A1 (en) | Method and multi-scale attention system for spatiotemporal change determination and object detection | |
| CN110363140A (en) | A kind of human action real-time identification method based on infrared image | |
| CN105631898B (en) | The infrared motion target detection method that conspicuousness merges when based on sky | |
| CN112734761B (en) | Industrial product image boundary contour extraction method | |
| CN109636758A (en) | A kind of floating on water object detecting method based on space-time dynamic operator | |
| CN106204594A (en) | A kind of direction detection method of dispersivity moving object based on video image | |
| CN109902618A (en) | A kind of sea ship recognition methods and device | |
| CN109934224A (en) | Small target detecting method based on markov random file and visual contrast mechanism | |
| CN111208479B (en) | Method for reducing false alarm probability in deep network detection | |
| CN110781721A (en) | Outdoor scene moving object detection method based on improved VIBE algorithm | |
| CN113111878B (en) | Infrared weak and small target detection method under complex background | |
| CN111667465A (en) | Metal hand basin defect detection method based on far infrared image | |
| CN110490903B (en) | Multi-target rapid capturing and tracking method in binocular vision measurement | |
| CN106056078A (en) | Crowd density estimation method based on multi-feature regression ensemble learning | |
| CN108960052A (en) | Ship overload detecting method based on video flowing | |
| CN109635679A (en) | A kind of real-time target sheet positioning and loop wire recognition methods | |
| CN110334703B (en) | Ship detection and identification method in day and night image | |
| CN108389191A (en) | The detection method in target shadow region in a kind of SAR image | |
| Fu et al. | Infrared small dim target detection under maritime near sea-sky line based on regional-division local contrast measure | |
| Selvi et al. | A novel approach for ship recognition using shape and texture | |
| CN111339824A (en) | Road surface sprinkled object detection method based on machine vision | |
| CN111089586A (en) | All-day star sensor star point extraction method based on multi-frame accumulation algorithm |
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 |