CN106372590A - Sea surface ship intelligent tracking system and method based on machine vision - Google Patents
Sea surface ship intelligent tracking system and method based on machine vision Download PDFInfo
- Publication number
- CN106372590A CN106372590A CN201610770608.1A CN201610770608A CN106372590A CN 106372590 A CN106372590 A CN 106372590A CN 201610770608 A CN201610770608 A CN 201610770608A CN 106372590 A CN106372590 A CN 106372590A
- Authority
- CN
- China
- Prior art keywords
- ship
- module
- target
- tracking
- machine vision
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Abstract
The present invention provides a sea surface ship intelligent tracking system and method based on machine vision. The system comprises an image collection module, a Mean-shift tracking module, a Kalman filter module, a multi-stage parallel detection module and online learning module. The sea surface ship intelligent tracking system and method based on the machine vision employs a Haar+AdaBoost combination algorithm to detect the sea surface ships, and the detection result is taken as the starting frame of the tracker so as to initialize the tracker and the detector and replace the manual delineation of a target area. The Kalman filter is configured for the detection and the tracking of the ship so as to shorten the detection range, reduce the calculation amount and improve the timeliness. Based on the Mean-shift tracking module, the stable feature, anti-shielding, simple calculation and insensitivity of target deformation, rotation and background movement; the multi-stage parallel detection module is employed to combine the set classifier based on the random forest, the Haar classifier and the nearest neighbor classifier so as to solve the difficulty that the ship tracking is failed caused by shielding and deformation and improve the tracking timeliness, the accuracy and the robustness.
Description
Technical field
The invention belongs to computer vision and field of artificial intelligence, it is related to sea ship intelligent-tracking technology, specifically
For a kind of sea ship intelligent tracking system based on machine vision and its method.
Background technology
The ocean socio-economic development national to has great strategic significance, especially wide to marine site, coastline
Very long country is even more important.Compared to land or sky field, sea unmanned vehicles at sea execute task and are faced with certain
Unique challenge a bit.Because sea environment is more complicated, in the mission completing environment detection, target recognition, automatic obstacle avoiding.Sea
Upper unmanned vehicles do not merely have to keep certain safe distance with land and it is necessary to course line on floating thing and fixture
Keep the distance of safety, this proposes higher requirement to the monitoring of environment and the identification of barrier.
During long-time tracking, tracked target will inevitably occur the side such as shape, illumination condition, yardstick
The change in face, also can produce situations such as block.Traditional tracking system, its front end needs to cooperate with detection module, works as inspection
After measuring tracked target, begin to enter tracking module, and hereafter, during detection module would not get involved in tracking.
But this system is with the presence of a critical defect: when tracked target change of shape or when blocking, follow the tracks of and be just easy to failure;Cause
, there is the tracking in the case of change of shape for long-time tracking or tracked target, prior art is using the side of detection in this
Method come to replace follow the tracks of.Although the method can improve tracking effect in some cases, it needs an offline study
Journey.I.e. before testing, need to select the sample of substantial amounts of tracked target to be learnt and train.This also implies that,
Training sample to cover tracked target it may happen that various deformation and various yardstick, attitudes vibration and illumination variation feelings
Condition.In other words, though the purpose of long-time tracking can be reached using the method for detection, the selection of its training sample is had very high
Require, otherwise it is difficult to ensure the robustness followed the tracks of.
Content of the invention
It is an object of the invention to overcoming the defect of prior art, provide a kind of sea ship intelligence based on machine vision
Tracking system and its method, in conjunction with machine vision technique and the present computer technology, it is possible to achieve long-term to ship target and
Stable tracking, possesses stronger adaptability, can solve the deformation scaling of target, the technical problem disappearing and reproducing, it is
The training sample of system can fully ensure that the robustness of tracking.
For solving the above-mentioned technical problem of prior art, a kind of sea ship intelligence based on machine vision of the present invention with
Track system is it is characterised in that include:
One image capture module, the described sea ship navigation chart picture of collection, by 1394 data circuits, will be described
Image information passes to a kalman filter module;
Described kalman filter module, the first detection to described ship target initialize, and then carry out
The filtering of boat trip image is estimated with ship position, and by described filtering with estimate after information be delivered separately to one
Mean-shift tracking module and a plural parallel stage detection module;
Described mean-shift tracking module, finds described target ship position, and by ship location dependent information
Pass to on-line study module;Described ship location dependent information includes: the position that ship target is located in each two field picture,
And ship clarification of objective value.
Described plural parallel stage detection module, carries out the detection of ship target, and passes to an on-line study module;
Described on-line study module, by the fusion to described image information, obtains the accurate positional information of ship, and
By described ship positional information be delivered separately to described mean-shift tracking module, kalman filter module and multistage simultaneously
Connection detection module is tracked the renewal of information.
Wherein, described image capture module, including ccd video camera, image pick-up card.Described kalman wave filter mould
Block includes haar grader and kalman wave filter.Described mean-shift tracking module includes: object module calculates submodule
Block, position candidate determination sub-module, candidate family calculating sub module, mean-shift iteration submodule and target location convergents
Module.Described plural parallel stage detection module includes: Ensemble classifier based on random forest, haar grader and arest neighbors divide
Class device.Described on-line study module includes: image co-registration submodule and p-n study submodule.
A kind of sea ship intelligent-tracking method based on machine vision of the present invention is it is characterised in that include following walking
Rapid:
(1) open video camera, gather video flowing, and be converted into digital picture, carry out data biography through 1394 data circuits
Pass;
(2) the ship target in the first frame of digital image is gone out by haar detection of classifier, as initial ship target, pass
Pass on-line study module;
(3) the initial ship target detecting with step (2), initializes kalman wave filter, determines ship target
Center, is filtered to each two field picture, and estimates the positional information of next frame ship;
(4) the initial ship target detecting with step (2), first initialization mean-shift tracking module and multistage
Detection module in parallel, then carries out target following and the detection of ship respectively;
(5) on-line study: the initial ship object initialization detecting with step (2), then learn to update by p-n
Based on mean-shift tracking module and plural parallel stage detection module;
(6) image co-registration: with determine follow the tracks of target positional information update kalman wave filter, thus improve ship with
The efficiency of track.
The detailed process of described step (2) is:
(21) detected using haar-like feature;
(22) using integrogram, haar-like feature evaluation is accelerated;
(23) adaboost Algorithm for Training is utilized to distinguish the strong classifier of ship and non-ship;
(24) using screening type cascade, strong classifier is cascaded to together;
(25) by strong classifier, target ship is identified, and ship with rectangular target circle out.
In described step (4), ship object module calculating is carried out based on mean-shift tracking module, its process is:
Kernel function fixed-bandwidth in described ship object module calculating is changed to the bandwidth of dynamic change, then to renewal
Object module afterwards is tracked;If the center of target area is x0, i.e. window center point vector value, if there being n pixel, useRepresent, object module has m eigenvalue, i.e. grey scale pixel value, and weight distribution depends on the band fat vector of kernel function
H, then the Multilayer networks of eigenvalue u=1 ... the m of object moduleFor:
In formula, It is normalization constants, h is the bandwidth of kernel function, and k is the pixel in core
Number;
When target is blocked, reliability near central point for the pixel will be significantly larger than the reliability away from central point
Property;δ[b(xi)-u] determine the attaching relation of pixel value and u-th eigenvalue, b (xi) it is gray value index function, if belonging to
This feature value is just 1, otherwise for 0;chThis constant factor ensure that
In the on-line study of described step (5), the picture containing ship target for the selection is positive sample, without ship target
Picture be negative sample.
Compared with prior art, the present invention includes advantages below and beneficial effect:
1. the present invention obtains ship target by the way of haar detector, need not manually select tracking target, improve
The efficiency that ship is followed the tracks of is so that whole tracking process is more intelligent.
2. the present invention adopts kalman wave filter, centered on the ship target followed the tracks of, chooses 2 times of tracking mesh in the picture
Target area is filtering space.On the one hand, carry out the filtering of image, improve the quality of image information;On the other hand, carry out image
In ship target location estimate, produce related tracking and detection target image.
3. the present invention has feature using mean-shift and stablizes, resists and block, calculate simplicity and to target deformation, rotation
The characteristic insensitive with background motion, using mean-shift tracking module, the kernel function during wherein ship object module calculates
Fixed-bandwidth is changed to the bandwidth of dynamic change, then the object module after updating is tracked, thus increased tracking can
By property and robustness.
4. the present invention adopts haar+adaboost combinational algorithm detection sea ship, testing result rising as tracker
Beginning frame, i.e. initialization tracker and detector, technosphere is replaced with this and sets the goal region.Pre- measurement of power by kalman wave filter
In ship detection being used in and having followed the tracks of, detection range is reduced with this, greatly less amount of calculation, improve real-time.
5. the present invention adopts plural parallel stage detection module, by the Ensemble classifier based on random forest, haar grader,
Nearest Neighbor Classifier gathers, and greatly in degree, solves the difficulty leading to ship to follow the tracks of failure because the factor such as blocking, deforming
Topic, further increases real-time, accuracy and the robustness of tracking.
6. the present invention can to marine vessel long-time real-time tracking, solve when ship deforms or seriously blocks
When, follow the tracks of the phenomenon of failure, there is precision and the real-time of very high tracking meanwhile.
Brief description
Fig. 1 is a kind of structured flowchart of embodiment based on the sea ship intelligent tracking system of machine vision for the present invention.
Fig. 2 is a kind of flow chart of embodiment based on the sea ship intelligent-tracking method of machine vision for the present invention.
Fig. 3 is a kind of workflow diagram of the plural parallel stage detection module of embodiment of the present invention.
Fig. 4 is a kind of workflow diagram of the mean-shift tracking module of embodiment of the present invention.
Fig. 5 is a kind of tracking result schematic diagram of embodiment of the present invention.Wherein, Fig. 5 a is the 3rd frame tracking result, Fig. 5 b
For the 21st frame tracking result, Fig. 5 c the 30th frame tracking result, Fig. 5 d is the 40th frame tracking result.
Specific embodiment
Below in conjunction with the accompanying drawings, the present invention is described in further details.
It is a kind of structural frames of embodiment based on the sea ship intelligent tracking system of machine vision for the present invention shown in Fig. 1
Figure.The system of the embodiment of the present invention includes: image capture module, mean-shift tracking module, kalman filter module, many
Level detection module in parallel, five parts of on-line study module.
Described image acquisition module, the described sea ship navigation chart picture of collection, by 1394 data circuits, will be described
Image information passes to a kalman filter module;
Described kalman filter module, the first detection to described ship target initialize, and then carry out
The filtering of boat trip image is estimated with ship position, and by described filtering with estimate after information be delivered separately to one
Mean-shift tracking module and a plural parallel stage detection module;Kalman wave filter is being tracked for sea ship
When, motion in the picture in the ship unit interval is considered as uniform motion, so the kinestate (position of a certain moment ship
With information such as velocity variations) can be described by kalman wave filter, the position of target is predicted further according to the equation of motion.One side
Face, kalman wave filter carries out the filtering of image, improves the quality of image;On the other hand, carry out image with kalman wave filter
In ship target location estimate, improve follow the tracks of efficiency.
Described mean-shift tracking module, finds described target ship position, and by ship location dependent information
Pass to on-line study module;Described ship location dependent information includes: the position that ship target is located in each two field picture,
And ship clarification of objective value.
Described plural parallel stage detection module, carries out the detection of ship target, and passes to an on-line study module;
Described on-line study module, by the fusion to described image information, obtains the accurate positional information of ship, and
Described ship positional information is delivered separately to described mean-shift tracking module, kalman wave filter and plural parallel stage inspection
Survey the renewal that module is tracked information.
Wherein, image capture module includes: ccd video camera, image pick-up card;Mean-shift tracking module includes following
Submodule: object module calculating, position candidate determination, candidate family calculating, mean-shift iteration, target location convergence.
Kalman filter module includes: haar grader and kalman wave filter.Plural parallel stage detection module includes: based on gloomy at random
The Ensemble classifier of woods, haar grader and nearest neighbor classifier.On-line study module includes: image co-registration, p-n study submodule
Block.
Fig. 2 show a kind of flow process of embodiment based on the sea ship intelligent-tracking method of machine vision for the present invention
Figure.This embodiment comprises the following steps:
(1) open video camera, gather video flowing, and be converted into digital picture, carry out data biography through 1394 data circuits
Pass;
(2) the ship target in the first frame of digital image is gone out by haar detection of classifier, as initial ship target, pass
Pass on-line study module;Step (2) obtains ship target by the way of haar detector, need not manually select tracking mesh
Mark, improves the efficiency of ship tracking so that whole tracking process is more intelligent.Its detailed process is:
(21) detected using haar-like feature;
(22) using integrogram, haar-like feature evaluation is accelerated;
(23) adaboost Algorithm for Training is utilized to distinguish the strong classifier of ship and non-ship;
(24) using screening type cascade, strong classifier is cascaded to together;
(25) by strong classifier, target ship is identified, and ship with rectangular target circle out.
(3) the initial ship target detecting with step (2), initializes kalman wave filter, determines ship target
Center, is filtered to each two field picture, and estimates the positional information of next frame ship;In this step (3), kalman filters
Ripple device, centered on the ship target followed the tracks of, the area choosing 2 times of tracking targets in the picture is filtering space.On the one hand, enter
The filtering of row image, improves the quality of image information;On the other hand, the ship target location carrying out in image is estimated, and produces phase
The tracking closed and detection target image.
(4) the initial ship target detecting with step (2), first initialization mean-shift tracking module and multistage
Detection module in parallel, then carries out target following and the detection of ship respectively.Its detailed process is:
Kernel function fixed-bandwidth in described ship object module calculating is changed to the bandwidth of dynamic change, then to renewal
Object module afterwards is tracked;Thus increased reliability and the robustness of tracking.If: the center of target area is x0, that is,
Window center point vector value, if there being n pixel, usesRepresent, object module has m eigenvalue, i.e. pixel grey scale
Value, weight distribution depends on the band fat vector h of kernel function, the then Multilayer networks of eigenvalue u=1 ... the m of object module
For:
In formula, It is normalization constants, h is the bandwidth of kernel function, and k is the pixel in core
Number;
When target is blocked, reliability near central point for the pixel will be significantly larger than the reliability away from central point
Property;δ[b(xi)-u] determine the attaching relation of pixel value and u-th eigenvalue, b (xi) it is gray value index function, if belonging to
This feature value is just 1, otherwise for 0;chThis constant factor ensure that
(5) on-line study: the initial ship object initialization detecting with step (2), then learn to update by p-n
Mean-shift tracker module and plural parallel stage detection module;
In the on-line study of this step (5), the picture containing ship target for the selection is positive sample, without ship target
Picture is negative sample.
(6) image co-registration: with determine follow the tracks of target positional information update kalman wave filter, thus improve ship with
The efficiency of track.
Claims (10)
1. a kind of sea ship intelligent tracking system based on machine vision is it is characterised in that include:
One image capture module, the described sea ship navigation chart picture of collection, by 1394 data circuits, by described image
Information transmission gives a kalman filtration module;
Described kalman filter module, the first detection to described ship target initialize, and then carry out ship
The filtering of navigation chart picture is estimated with ship position, and by described filtering with estimate after information be delivered separately to one
Mean-shift tracking module and a plural parallel stage detection module;
Described mean-shift tracking module, finds described target ship position, and ship location dependent information is transmitted
To on-line study module;Described ship location dependent information includes: the position that ship target is located in each two field picture, and
Ship clarification of objective value;
Described plural parallel stage detection module, carries out the detection of ship target, and passes to an on-line study module;
Described on-line study module, by the fusion to described image information, obtains the accurate positional information of ship, and by institute
State ship positional information and be delivered separately to described mean-shift tracking module, kalman filter module and plural parallel stage inspection
Survey the renewal that module is tracked information.
2. a kind of sea ship intelligent tracking system based on machine vision according to claim 1 it is characterised in that: institute
The image capture module stated, including ccd video camera, image pick-up card.
3. a kind of sea ship intelligent tracking system based on machine vision according to claim 1 it is characterised in that: institute
The kalman filter module stated includes haar grader and kalman wave filter.
4. a kind of sea ship intelligent tracking system based on machine vision according to claim 1 it is characterised in that: institute
The mean-shift tracking module stated includes: object module calculating sub module, position candidate determination sub-module, candidate family calculate
Submodule, mean-shift iteration submodule and target location convergence submodule.
5. a kind of sea ship intelligent tracking system based on machine vision according to claim 1 it is characterised in that: institute
The plural parallel stage detection module stated includes: Ensemble classifier based on random forest, haar grader and nearest neighbor classifier.
6. a kind of sea ship intelligent tracking system based on machine vision according to claim 1 it is characterised in that: institute
The on-line study module stated includes: image co-registration submodule and p-n study submodule.
7. system described in a kind of any one using claim 1 to 6 come to realize sea ship intelligence based on machine vision with
Track method is it is characterised in that comprise the following steps:
(1) open video camera, gather video flowing, and be converted into digital picture, carry out data transfer through 1394 data circuits;
(2) the ship target in the first frame of digital image is gone out by haar detection of classifier, as initial ship target, pass to
On-line study module;
(3) the initial ship target detecting with step (2), initializes kalman wave filter, determines the center of ship target
Position, is filtered to each two field picture, and estimates the positional information of next frame ship;
(4) the initial ship target detecting with step (2), first initialization mean-shift tracking module and plural parallel stage
Detection module, then carries out target following and the detection of ship respectively;
(5) on-line study: the initial ship object initialization detecting with step (2), then learn to update by p-n
Mean-shift tracker module and plural parallel stage detection module;
(6) image co-registration: update kalman wave filter with the positional information determining tracking target, thus improving ship tracking
Efficiency.
8. a kind of sea ship intelligent-tracking method based on machine vision according to claim 7 is it is characterised in that institute
The detailed process of the step (2) stated is:
(21) detected using haar-like feature;
(22) using integrogram, haar-like feature evaluation is accelerated;
(23) adaboost Algorithm for Training is utilized to distinguish the strong classifier of ship and non-ship;
(24) using screening type cascade, strong classifier is cascaded to together;
(25) by strong classifier, target ship is identified, and ship with rectangular target circle out.
9. a kind of sea ship intelligent-tracking method based on machine vision according to claim 7 it is characterised in that
In described step (4), ship object module calculating is carried out based on mean-shift tracking module, its process is:
Kernel function fixed-bandwidth during described ship object module calculates is changed to the bandwidth of dynamic change, then to updating after
Object module is tracked;If the center of target area is x0, i.e. window center point vector value, if there being n pixel, use { xi}i=
1 ... n represents, object module has m eigenvalue, i.e. grey scale pixel value, and weight distribution depends on the band fat vector h of kernel function,
The then Multilayer networks of eigenvalue u=1 ... the m of object moduleFor:
In formula, It is normalization constants, h is the bandwidth of kernel function, and k is the pixel in core
Number;
When target is blocked, reliability near central point for the pixel will be significantly larger than the reliability away from central point;δ[b
(xi)-u] determine the attaching relation of pixel value and u-th eigenvalue, b (xi) it is gray value index function, if belonging to this feature
Value just for 1, otherwise for 0;chThis constant factor ensure that
10. a kind of sea ship intelligent-tracking method based on machine vision according to claim 7 it is characterised in that
In the on-line study of described step (5), the picture containing ship target for the selection is positive sample, and the picture without ship target is
Negative sample.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610770608.1A CN106372590A (en) | 2016-08-29 | 2016-08-29 | Sea surface ship intelligent tracking system and method based on machine vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610770608.1A CN106372590A (en) | 2016-08-29 | 2016-08-29 | Sea surface ship intelligent tracking system and method based on machine vision |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106372590A true CN106372590A (en) | 2017-02-01 |
Family
ID=57901972
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610770608.1A Pending CN106372590A (en) | 2016-08-29 | 2016-08-29 | Sea surface ship intelligent tracking system and method based on machine vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106372590A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107197199A (en) * | 2017-05-22 | 2017-09-22 | 哈尔滨工程大学 | A kind of intelligent monitoring and controlling device and method for tracking target |
CN107292917A (en) * | 2017-06-15 | 2017-10-24 | 武汉理工大学 | A kind of Dragging recognition methods based on Computer Vision |
CN108471497A (en) * | 2018-03-02 | 2018-08-31 | 天津市亚安科技有限公司 | A kind of ship target real-time detection method based on monopod video camera |
WO2019101220A1 (en) * | 2017-12-11 | 2019-05-31 | 珠海大横琴科技发展有限公司 | Deep learning network and average drift-based automatic vessel tracking method and system |
CN110097101A (en) * | 2019-04-19 | 2019-08-06 | 大连海事大学 | A kind of remote sensing image fusion and seashore method of tape sorting based on improvement reliability factor |
CN116246205A (en) * | 2023-02-27 | 2023-06-09 | 杭州数尔安防科技股份有限公司 | Optimization method and system for ship tracking and character recognition |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101916446A (en) * | 2010-07-23 | 2010-12-15 | 北京航空航天大学 | Gray level target tracking algorithm based on marginal information and mean shift |
CN102136076A (en) * | 2011-03-14 | 2011-07-27 | 徐州中矿大华洋通信设备有限公司 | Method for positioning and tracing underground personnel of coal mine based on safety helmet detection |
CN103886322A (en) * | 2012-12-20 | 2014-06-25 | 中山大学深圳研究院 | Video target tracking method based on SVM and Mean-Shift |
CN105426928A (en) * | 2014-09-19 | 2016-03-23 | 无锡慧眼电子科技有限公司 | Pedestrian detection method based on Haar characteristic and EOH characteristic |
CN105893957A (en) * | 2016-03-30 | 2016-08-24 | 上海交通大学 | Method for recognizing and tracking ships on lake surface on the basis of vision |
-
2016
- 2016-08-29 CN CN201610770608.1A patent/CN106372590A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101916446A (en) * | 2010-07-23 | 2010-12-15 | 北京航空航天大学 | Gray level target tracking algorithm based on marginal information and mean shift |
CN102136076A (en) * | 2011-03-14 | 2011-07-27 | 徐州中矿大华洋通信设备有限公司 | Method for positioning and tracing underground personnel of coal mine based on safety helmet detection |
CN103886322A (en) * | 2012-12-20 | 2014-06-25 | 中山大学深圳研究院 | Video target tracking method based on SVM and Mean-Shift |
CN105426928A (en) * | 2014-09-19 | 2016-03-23 | 无锡慧眼电子科技有限公司 | Pedestrian detection method based on Haar characteristic and EOH characteristic |
CN105893957A (en) * | 2016-03-30 | 2016-08-24 | 上海交通大学 | Method for recognizing and tracking ships on lake surface on the basis of vision |
Non-Patent Citations (2)
Title |
---|
徐光柱等著: "《实用性目标检测与跟踪算法原理与应用》", 30 June 2015, 国防工业出版 * |
杨杰等著: "《视频目标检测和跟踪及其应用》", 31 August 2012, 上海交通大学出版社 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107197199A (en) * | 2017-05-22 | 2017-09-22 | 哈尔滨工程大学 | A kind of intelligent monitoring and controlling device and method for tracking target |
CN107292917A (en) * | 2017-06-15 | 2017-10-24 | 武汉理工大学 | A kind of Dragging recognition methods based on Computer Vision |
WO2019101220A1 (en) * | 2017-12-11 | 2019-05-31 | 珠海大横琴科技发展有限公司 | Deep learning network and average drift-based automatic vessel tracking method and system |
US10706285B2 (en) | 2017-12-11 | 2020-07-07 | Zhuhai Da Hengqin Technology Development Co., Ltd. | Automatic ship tracking method and system based on deep learning network and mean shift |
CN108471497A (en) * | 2018-03-02 | 2018-08-31 | 天津市亚安科技有限公司 | A kind of ship target real-time detection method based on monopod video camera |
CN110097101A (en) * | 2019-04-19 | 2019-08-06 | 大连海事大学 | A kind of remote sensing image fusion and seashore method of tape sorting based on improvement reliability factor |
CN110097101B (en) * | 2019-04-19 | 2022-09-13 | 大连海事大学 | Remote sensing image fusion and coastal zone classification method based on improved reliability factor |
CN116246205A (en) * | 2023-02-27 | 2023-06-09 | 杭州数尔安防科技股份有限公司 | Optimization method and system for ship tracking and character recognition |
CN116246205B (en) * | 2023-02-27 | 2024-04-19 | 杭州数尔安防科技股份有限公司 | Optimization method and system for ship tracking and character recognition |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106372590A (en) | Sea surface ship intelligent tracking system and method based on machine vision | |
US20220197281A1 (en) | Intelligent decision-making method and system for unmanned surface vehicle | |
CN104200495B (en) | A kind of multi-object tracking method in video monitoring | |
CN103149939B (en) | A kind of unmanned plane dynamic target tracking of view-based access control model and localization method | |
CN110889324A (en) | Thermal infrared image target identification method based on YOLO V3 terminal-oriented guidance | |
CN103426179B (en) | A kind of method for tracking target based on mean shift multiple features fusion and device | |
CN109657541A (en) | A kind of ship detecting method in unmanned plane image based on deep learning | |
CN111091095B (en) | Method for detecting ship target in remote sensing image | |
CN103336947B (en) | Based on conspicuousness and structural infrared moving small target recognition methods | |
CN103164858A (en) | Adhered crowd segmenting and tracking methods based on superpixel and graph model | |
CN104318258A (en) | Time domain fuzzy and kalman filter-based lane detection method | |
CN110275153A (en) | A kind of waterborne target detection and tracking based on laser radar | |
CN105224947A (en) | Sorter training method and system | |
CN106875421A (en) | A kind of multi-object tracking method and device | |
CN105160649A (en) | Multi-target tracking method and system based on kernel function unsupervised clustering | |
CN111723632B (en) | Ship tracking method and system based on twin network | |
CN105427342A (en) | Method and system for detecting and tracking underwater small-target sonar image target | |
CN105654516A (en) | Method for detecting small moving object on ground on basis of satellite image with target significance | |
CN111968046A (en) | Radar photoelectric sensor target association fusion method based on topological structure | |
CN107798691A (en) | A kind of unmanned plane independent landing terrestrial reference real-time detecting and tracking method of view-based access control model | |
CN104156976A (en) | Multiple characteristic point tracking method for detecting shielded object | |
CN114639064B (en) | Water level identification method and device | |
CN104182992A (en) | Method for detecting small targets on the sea on the basis of panoramic vision | |
CN110490903B (en) | Multi-target rapid capturing and tracking method in binocular vision measurement | |
CN110427030A (en) | A kind of unmanned boat based on Tiny-YOLOship algorithm of target detection independently docks recovery method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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: 20170201 |