CN101145200A - Inner river ship automatic identification system of multiple vision sensor information fusion - Google Patents

Abstract

An automatic identification system, which is integrated by the information of a multi-visual sensor and is used on inland river ships, comprises a large-scale monitoring visual sensor which can be used on monitor fairways, an express-ball visual sensor which can be used for shooting close-up images of ship bodies and name plates of ships, and a microprocessor which can be used for tracing target ships, identifying images and summarizing traffic situations of inland rivers; the large-scale monitoring visual sensor can realize multi-target tracing for ships on fairways; when a ship enters the monitoring area, the system automatically produces an event and an ID of the ship to control the rotating and focusing of the express-ball visual sensor; the express-ball visual sensor focuses on the traced vessel to shoot; the height of the vessel body which is above water and the load can be estimated by detecting of the outline of the close-up image; and at the same time, by positioning the ship cockpit, the number of the name plate can be shot and identified; by integrating multi-visual sensors and automatically collecting the basic data of inland river traffic by a computer, the present invention can effectively manage inland rivers.

Description

The inner river ship automatic identification system that multiple vision sensor information merges

Technical field

The invention belongs to ship automatic identification system, relate in particular to the tonnage video detection technology of multiple vision sensor information integration technology, multiple target tracking technology, framing candid photograph technology, image recognition technology, ship and database and computer networking technology in the boats and ships application of identification automatically.

Background technology

Ship automatic identification system (AIS) promptly is a kind of novel navigational aid peculiar to vessel, simultaneously also is the gordian technique in a kind of monitoring management system of ship's navigation.Its function and performance standard are formulated by International Maritime Organization (IMO) (IMO), International Association of Lighthouse Authorities (IALA), the joint research of international organizations such as (ITU) of International Telecommunication Association.Its fundamental purpose is by more accurate, detailed navigation environment information being provided for the Ship Controling personnel, enabling to monitor and control ship's navigation better, strengthening sea going security.International Maritime Organization (IMO) (IMO) regulation, on July 1,1 day to 2008 July in 2002 navigation is navigated by water boats and ships more than domestic air route 500 gross tons in the boats and ships more than international airline 300 gross tons and Convention country, carries out ship automatic identification system AIS (AutomaticIdentification System) equipment that is equipped with stage by stage.The appearance of AIS with and widespread use, not only improved sea going safety coefficient, also provide a kind of brand-new and reliable method for marine ships traffic flow investigation.Be used for static state that information that the technology of boats and ships identification aspect mainly contains visual observation, radar observation, aeroplane photography and utilize onboard AIS to broadcast carries out and multidate information at present and four kinds of methods such as discern.

Also having some scholars to adopt the RFID technology in recent years is unlimited REID is discerned boats and ships at narrow waters identity, the RFID technology is applied to ship monitor management, the management of boats and ships entrance and clearance, dangerous material ship's fix management, ship overload management, checks a series of ports aviation management such as expropriation and management reason and manage project, but the RFID technology also has certain limitation: 1) each boats and ships all needs to be equipped with a rfid card, has increased the use cost of every boats and ships; 2) the radio-frequency (RF) identification communication distance is subjected to certain restriction, implements very difficulty for the navigation channel in wide waters; 3) rfid card that is used to discern watercraft identification all is to belong to long distance identification, must adopt active rfid card, therefore need constantly power to rfid card.

Traffic on inland waters observation procedure and maritime traffic observation procedure have a lot of something in common, because several years ago competent authorities are little to dropping in the traffic administration of inland river, the information system management level is not high, caused traffic on inland waters detection means aspect to lag behind the situation that maritime traffic detects relatively, along with country dashed forward to the attention and the showing of the critical role of inland water transport in national product of inland water transport, the importance of developing a kind of ship automatic identification system has obtained insider's common concern.

Inland water transport is an ingredient important in the multimodal transport network, particularly has irreplaceable advantage and effect in the transportation of emphasis bulk supply tariff.As a kind of occupation of land less, pollution is little, energy consumption is little, freight volume is big means of transportation.Particularly under the situation that the fast-developing process of Building Trade in China is accelerated, traffic above-ground resource day is becoming tight and energy prices skyrocket, the advantage of inland water transport meets the requirement of sustainable development to transportation, and can actively promote the waterway dredging regulation, take in comprehensive benefits such as shipping, flood control, generating and irrigation concurrently, promote the basin expanding economy; Its politics to country, military affairs and safety are significant, can bring into play special effect under the emergency circumstance such as rescue and relief work; Its development is implementing important effect to the policy of sustainable development.

Inland water transport has advantages such as freight rate is low, freight volume is big, energy consumption is low, reduced investment, is an important component part of China, our province transportation.The reasonable cooperation of navigation channel and ship type is an importance of building inland navigation projects.The accurate statistics of the inland water transport tonnage of ship can provide complete, decision-making foundation accurately to relevant department.But the measurement of China's inland water transport tonnage of ship statistics mainly is the method that relies on manual observation, guestimate at present, and efficient is low, the data error of statistics is big and have the careless omission phenomenon.Therefore be badly in need of the real-time collection and the intelligent measure technology of the processing and the shipping tonnage of ship of development traffic on inland waters stream information, the realization automated job is to raise the efficiency the order of accuarcy with statistics.This is the problem of a sensor measurement and information processing from essence.

2006, marine board of the People's Republic of China (PRC) has proposed all must use the unified name of vessel sign board or the requirement of name of vessel lamp box to national spitkit (200 gross tons are following), these automatic identifications that are defined as name of vessel provide better basis, comparative maturity aspect the vehicle license automatic identification technology at present, some has dropped into commercial the application.In the notice about issue " spitkit name of vessel flag management tentative method " in [2006] No. 594 files of seagoing vessel oceangoing ship of marine board of the People's Republic of China (PRC) relevant regulation is arranged.

In present information processing and Measurement and Control System, mostly need to use sensor as the means of obtaining information.Vision is human important means of observing the world and the cognitive world.According to statistics, the human information that obtains from the external world has 80% to be obtained by vision approximately.This had both shown containing much information of vision, and human have higher utilization factor to visual information, also embodied the importance of visual performance simultaneously.In this sense, vision has not only comprised the perception of light signal, also comprises the overall process of obtaining, transmit, handling and understand to visual information.

Current, the developing rapidly and combination of embedded calculating, radio communication and sensor technology, made people can almost be gather ubiquitously, the data of transmission and store video and audio frequency.If the data of these magnanimity can obtain in time and accurately analyzing and understanding, will play a significant role in cruiseway monitoring field.At present video detecting device has been installed in some developed regions of China on the main trunk road of inland river, but these video datas only are used for afterwards affirmation, also lack profound analytical applications.Dynamic image is understood technology provides powerful technical support for a large amount of video data processing, analysis.Particularly use at the aspects such as real-time collection, processing, analysis and even decision-making management of cruiseway telecommunication flow information and will demonstrate very strong advantage.

Therefore the realization of real-time collection, processing, analysis and the various intelligent processing methods of cruiseway video traffic stream information, no matter to traffic control, traffic administration, traffic programme, waterway network construction, the supervision of law enforcement in inland river, still the realization to following cruiseway intelligent transportation system function all has important significance for theories and practical value.

Summary of the invention

Mainly need to rely on artificial 24 hours and do not stop and observe continuously in order to overcome existing inland navigation craft traffic flow visual observation method, the observation procedure estimated value is rough, observed efficiency is low, the statistics error big, detect deficiencies such as careless omission occur, the invention provides a kind of master data of collecting traffic on inland waters automatically, the inner river ship automatic identification system that can carry out the multiple vision sensor information fusion of effective traffic on inland waters management by many visual sensings fusions and computing machine.

The technical solution adopted for the present invention to solve the technical problems is:

The inner river ship automatic identification system that a kind of multiple vision sensor information merges comprises the large-range monitoring vision sensor that monitors the navigation channel, carries out feature and capture the fire ball vision sensor of hull image and name of vessel sign board image and be used for the microprocessor that boats and ships are added up image tracing, image recognition and traffic on inland waters situation; Described large-range monitoring vision sensor, fire ball vision sensor are connected with microprocessor, described microprocessor comprises: image-display units is used for showing the video image of whole cruiseway, the general image of tracking boats and ships and the driving cabin image of following the tracks of boats and ships;

Large-range monitoring vision sensor calibration module is used to set up the navigation channel chart picture in space and the corresponding relation of the video image that is obtained;

The video data Fusion Module is used to control the rotation and the focusing of fire ball vision sensor, makes the fire ball vision sensor can aim at the tracking boats and ships and carries out the feature candid photograph;

The dummy line customized module, be used to customize the detection dummy line on the monitoring navigation channel, article three, detect that dummy line equates at interval and perpendicular to the navigation channel, article three, detect dummy line in the field range of large-range monitoring vision sensor, intersect a middle installation site of detecting dummy line and fire ball vision sensor;

Boats and ships enter and detect the dummy line detection module, are used for entering the detection dummy line if any boats and ships, produce an incident automatically, call corresponding processing module during incident of every generation automatically;

Boats and ships ID number and deposit the automatically-generating module of following the tracks of the boats and ships image folder, be used for the boats and ships of the dummy line of the visual field ragged edge that just enters the large-range monitoring vision sensor are named, and generate a file simultaneously with ID number name of these boats and ships, be used to deposit the close-up image of these boats and ships;

Module is captured in boats and ships general image location, is used to locate the image of the integral body of capturing these boats and ships;

The boats and ships profile detection module, be used to detect the edge contour of these boats and ships and estimate the height that this boats and ships hull surfaces, the notch cuttype edge detection algorithm of employing optimum-be that the canny edge detection algorithm carries out rim detection to the close-up image of this hull, obtain hull load-carrying feature and driving cabin characteristic information;

The multiple target tracking module is used to follow the tracks of the passing boats and ships on the navigation channel, the pixel of prospect boats and ships is extracted in the navigation channel background at the low-level image feature layer, and this tracking module comprises:

The adaptive background reduction unit, be used for cutting apart in real time dynamic object, the feature of using mixture gaussian modelling to come each pixel in the phenogram picture frame, when obtaining new picture frame, upgrade mixture gaussian modelling, on each time period, select the subclass of mixture gaussian modelling to characterize current background, if the pixel of present image and mixture gaussian modelling are complementary, judge that then this point is a background dot, otherwise judge that this point is the foreground point;

The connected region identify unit is used to extract prospect boats and ships object, adopts eight connected region extraction algorithms, utilize corrosion and expansion operator to remove isolated noise foreground point and the aperture of filling up the target area respectively, testing result is designated as { Rei, i=1,2,3 ... n}, the connected region that will detect gained at last projects on the initial prospect point set F again, gets communication with detection result { Ri=Rei ∩ F to the end, i=1,2,3,, n};

Tracking cell, be used for utilizing the color characteristic of ship target object to find the position and the size at motion ship target object place at video image, in the next frame video image, with moving target current position and big or small initialization search window, repeat this process and just can realize Continuous Tracking target;

Module is captured in location, ship-handling cabin, be used to locate the approximate location of the suspension name of vessel sign board of these boats and ships or lamp box and this position captured the image of the name of vessel sign trade mark, when tracked ship target object process detects dummy line, computing machine sends a series of instruction control fire ball vision sensors 2 automatically and turns to the residing orientation of ship target object, good focal length is captured then, the image of capturing has three at least, wherein an image is positive side position, and other two images are looked side ways the position about being respectively;

Name of vessel sign trade mark identification module is used to discern the name of vessel sign trade mark of shippping traffic, and this identification module comprises:

Based on the positioning unit of the name of vessel sign board of color model, be used for determining the position of name of vessel sign board at hull; Employing comes station keeping ship name sign board based on the LUV color model and the edge contour algorithm of even color space; Whether at first slightly judge has blueness or white to exist, if exist comparatively significantly blue portion or comparatively significantly white portion, jointing edge profile algorithm is determined the position of name of vessel sign board in the image of being captured;

The extraction unit of name of vessel sign board is used for the name of vessel sign board is extracted separately from hull, adopts the relation of complementary colors to extract the name of vessel sign board;

The correcting unit of name of vessel sign board, be used for that institute is extracted the rectification of name of vessel sign board image and be the front elevation picture, by the hough conversion rectangular coordinate is transformed into the parameter coordinate system, utilize K average fitting a straight line algorithm, detect two angles of ship board, utilize the method for " rotation+mistake is cut " conversion again, promptly earlier the angle of inclination is rotated and does to correct for the first time, carry out wrong shearization again and do rectification for the second time, deformation angle is carried out stretching conversion, whole licence plate is corrected into the licence plate of facing of a standard;

The Character segmentation unit of name of vessel sign board is used for the character on the name of vessel sign board after correcting is cut apart; Adopt the masterplate coupling each character on the name of vessel sign board to be cut apart in conjunction with the vertical and horizontal sciagraphy;

Character recognition unit, the character of each name of vessel sign board after being used for cutting apart is discerned, at first to carry out the character normalized, character is normalized to unified size, adopt the recognition methods of sorter for Chinese Character Recognition, the Chinese character after the normalized and standard Chinese character are mated discern; Adopt recognition methods for English alphabet, arabic numeral identification based on the BP neural network;

Boats and ships essential information retrieval module, being used for the name of vessel sign trade mark is major key, inquiry boats and ships essential information in the boats and ships Basic Information Table.

As preferred a kind of scheme: described large-range monitoring vision sensor is the wide-angle vision sensor, is installed in a side of cruiseway.

Or: described large-range monitoring vision sensor is an omnibearing vision sensor, described omnibearing vision sensor is installed in the centre in navigation channel, described omnibearing vision sensor comprises the evagination catadioptric minute surface that is used for reflecting field, navigation channel object, evagination catadioptric minute surface down, be used to prevent anaclasis and the saturated dark circles cone of light, the dark circles cone is fixed on the center of catadioptric minute surface male part, be used to support the transparent cylinder of evagination catadioptric minute surface, be used to take the camera of imaging body on the evagination mirror surface, camera facing to the evagination mirror surface up.

Further, described microprocessor also comprises: shipping draft and tonnage estimation block, be used to estimate the dead weight capacity of current boats and ships, the height that surfaces according to this boats and ships hull and leave long, high, wide data of these boats and ships in the boats and ships master database in, extrapolate this shipping draft, then the estimation dead weight capacity that calculates these boats and ships from the length and the width data of this ship; Estimation equation is: ship side is exposed at the height on the water surface during draft during Ship's Cargo draft=boats and ships ship side overall height-boats and ships hole capital after selling all securities-boats and ships loading; Water discharge (the metric ton)=wide * goods of long * draft * side's mode coefficient (cubic meter)/0.9756 (seawater) or 1 (fresh water) (cubic meter), displacement tonnage is as the deadweight tonnage of estimation boats and ships.

Further again, described microprocessor also comprises: the speed of the ship in metres per second detecting unit, be used to detect the headway of boats and ships, when boats and ships pass through to detect dummy line I respectively, when II and III, setting boats and ships ID is ShipID, initial tracing positional is StartPosition, initial time is StartTime, the end tracing positional is EndPosition, concluding time is EndTime), boats and ships ID enters at boats and ships and generates automatically when one of outermost detects dummy line, initial tracing positional is to enter position of detecting dummy line of outermost, initial time is to enter system time that detects dummy line of outermost, finishing tracing positional is to leave the monitoring field just to have run into position of detecting dummy line of outermost, concluding time is to leave the monitoring field just to have run into time of detecting dummy line of outermost, comes the headway of Ship ' by following formula:

$V\left(\mathrm{ShipID}\right)=\frac{\mathrm{EndPosition}-\mathrm{StartPosition}}{\mathrm{EndTime}-\mathrm{StartTime}}---\left(5\right).$

Further, described microprocessor also comprises: boats and ships up-downgoing judging unit, be used to judge the navigation direction of boats and ships, if what the boats and ships object of following the tracks of was at first run into is to detect dummy line I to be defined as up direction, be to detect dummy line III to be defined as down direction if the boats and ships object of following the tracks of at first runs into.

Described microprocessor also comprises: shippping traffic record generation module, be used for generating automatically a shippping traffic record, thereby traffic on inland waters amount statistics etc. performs data and prepares, and boat note comprises: boats and ships ID, the name of vessel sign trade mark, captain, the beam, height, estimation tonnage, detection time, speed of the ship in metres per second and up-downlink direction.

Described microprocessor also comprises: volume of traffic statistical module is used to add up hourly traffic volume, daily traffic volume, the monthly average volume of traffic or the annual volume of traffic by certain point on the navigation channel; The volume of traffic is meant the boats and ships number that passes through a certain waters in the unit interval, and the size of the volume of traffic directly reflects the busy extent of this waters vessel traffic; Because the volume of traffic is time dependent, so adopt hourly traffic volume, daily traffic volume, the monthly average volume of traffic or the annual volume of traffic to represent;

Hour the volume of traffic with formula (13) expression:

$\stackrel{\‾}{Q}(i{)}_{\mathrm{up}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\Σ}}}{Q}_{\mathrm{up}}$ (13)

$\stackrel{\‾}{Q}(i{)}_{\mathrm{down}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\Σ}}}{Q}_{\mathrm{down}}$

$\stackrel{\_}{Q}\left(i\right)=\stackrel{\_}{Q}(i{)}_{\mathrm{up}}+\stackrel{\_}{Q}(i{)}_{\mathrm{down}}$

In the formula: i is the integer between 0～23,, Q (i) _UpBe up boats and ships number in certain i hour, Q (i) _DownBe descending boats and ships number in certain i hour, Q (i) is the boats and ships number of up-downgoing in certain i hour;

Daily traffic volume is represented with formula (14):

$\stackrel{\_}{Q}(d{)}_{\mathrm{up}}=\underset{i=0}{\overset{i=23}{\mathrm{\Σ}}}Q(i{)}_{\mathrm{up}}$ (14)

$\stackrel{\_}{Q}(d{)}_{\mathrm{duwn}}=\underset{i=0}{\overset{i=23}{\mathrm{\Σ}}}Q(i{)}_{\mathrm{down}}$

$\stackrel{\_}{Q}\left(d\right)=\stackrel{\_}{Q}(d{)}_{\mathrm{up}}+\stackrel{\_}{Q}(d{)}_{\mathrm{down}}$

In the formula: Q (d) _UpBe up boats and ships number in certain day, Q (d) _DownFor descending boats and ships in certain day are counted the boats and ships number of Q (d) for up-downgoing in certain day;

The statistics monthly average volume of traffic or the annual volume of traffic adds up as long as the boats and ships number of up-downgoing every day in certain month, certain year added up.

Described microprocessor also comprises: the deadweight tonnage statistical module is used to add up hour deadweight tonnage, day deadweight capacity position, monthly average deadweight tonnage or annual deadweight tonnage by certain point on the navigation channel; Count hour deadweight tonnage by certain point on the navigation channel, day deadweight capacity position, monthly average day deadweight capacity position or annual deadweight tonnage;

Hour deadweight tonnage with formula (15) expression,

$\stackrel{\_}{W}(i{)}_{\mathrm{up}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\Σ}}}{W}_{\mathrm{up}}$ (15)

$\stackrel{\_}{W}(i{)}_{\mathrm{down}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\Σ}}}{w}_{\mathrm{down}}$

$\stackrel{\_}{W}\left(i\right)=\stackrel{\_}{W}(i{)}_{\mathrm{up}}+\stackrel{\_}{W}(i{)}_{\mathrm{down}}$

In the formula: i is the integer between 0～23,, W (i) _UpBe up deadweight tonnage summation in certain i hour, W (i) _DownBe descending deadweight tonnage summation in certain i hour, W (i) is the deadweight tonnage summation of up-downgoing in certain i hour;

Day deadweight capacity position can use formula (16) to represent,

$\stackrel{\_}{W}(d{)}_{\mathrm{up}}=\underset{i=0}{\overset{i=23}{\mathrm{\Σ}}}W(i{)}_{\mathrm{up}}$ (16)

$\stackrel{\_}{W}(d{)}_{\mathrm{down}}=\underset{i=0}{\overset{i=23}{\mathrm{\Σ}}}W(i{)}_{\mathrm{down}}$

$\stackrel{\_}{W}\left(d\right)=\stackrel{\_}{W}(d{)}_{\mathrm{up}}+\stackrel{\_}{W}(d{)}_{\mathrm{down}}$

In the formula: W (d) _UpBe up boats and ships deadweight tonnage summation in certain day, W (d) _DownBe descending boats and ships deadweight tonnage summation in certain day, W (d) is the boats and ships deadweight tonnage summation of up-downgoing in certain day;

Statistics monthly average boats and ships deadweight tonnage summation or annual boats and ships deadweight tonnage summation adds up as long as the boats and ships deadweight tonnage summation of up-downgoing every day in certain month, certain year added up.

Described microprocessor also comprises: the traffic density statistical module is used to add up hour traffic density, daily traffic density, month traffic density or year traffic density by certain point on the navigation channel; Traffic density is meant a boats and ships number that travels in the unit length waters, it reflects the dense degree of boats and ships in this waters, and traffic density is represented with formula (17).

$\mathrm{\ρ}=\frac{N}{\mathrm{WL}}---\left(17\right)$

In the formula: ρ is the traffic density of certain moment; N is the boats and ships number; L is the length in observation navigation channel; W is the width in observation navigation channel.

Relational expression between the parameters such as the volume of traffic, traffic density, speed is called traffic flow model, formula (18) expression of traffic flow basic model,

$\mathrm{\ρ}=\frac{Q}{\mathrm{WV}}---\left(18\right)$

In the formula: ρ is a traffic density; Q is the volume of traffic; V is a speed; W is the width in observation navigation channel.Ship's navigation average velocity hourly is represented with formula (19):

$\stackrel{\_}{V}(i{)}_{\mathrm{up}}=\frac{1}{N}\underset{n=i}{\overset{n=i+1}{\mathrm{\Σ}}}{V}_{\mathrm{up}}$ (19)

$\stackrel{\_}{V}(i{)}_{\mathrm{down}}=\frac{1}{N}\underset{n=i}{\overset{n=i+1}{\mathrm{\Σ}}}{V}_{\mathrm{down}}$

$\stackrel{\_}{V}\left(i\right)=\frac{1}{2}\left(\stackrel{\_}{V}(i{)}_{\mathrm{up}}+\stackrel{\_}{V}(i{)}_{\mathrm{down}}\right)$

In the formula: i is the integer between 0～23,, V (i) _UpThe average speed per hour of up boats and ships in certain i hour, V (i) _DownBe the average speed per hour of descending boats and ships in certain i hour, V (i) is the average speed per hour of up-downgoing boats and ships in certain i hour; The traffic density that we can calculate hour according to formula (18), computing formula is provided by (20),

$\mathrm{\ρ}(i{)}_{\mathrm{up}}=\frac{\stackrel{\_}{Q}(i{)}_{\mathrm{up}}}{\stackrel{\‾}{W*V}(i{)}_{\mathrm{up}}}$ (20)

$\mathrm{\ρ}(i{)}_{\mathrm{down}}=\frac{\stackrel{\_}{Q}(i{)}_{\mathrm{down}}}{\stackrel{\‾}{W*V}(i{)}_{\mathrm{down}}}$

$\mathrm{\ρ}\left(i\right)=\frac{\stackrel{\_}{Q}\left(i\right)}{\stackrel{\‾}{W*V}\left(i\right)}$

In the formula: i is the integer between 0～23,, ρ (i) _UpBe up navigation channel traffic density in certain i hour, ρ (i) _DownBe descending navigation channel traffic density in certain i hour, ρ (i) is certain i hour inner passage traffic density;

In a certain waters, channel span is fixing, and after definite magnitude of traffic flow and these two parameters of headway, application of formula (20) is extrapolated traffic density;

Daily traffic density calculates with formula (21) expression,

$\stackrel{\_}{\mathrm{\ρ}}(d{)}_{\mathrm{up}}=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\Σ}}}\mathrm{\ρ}(i{)}_{\mathrm{up}}$ (21)

$\stackrel{\_}{\mathrm{\ρ}}(d{)}_{\mathrm{down}}=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\Σ}}}\mathrm{\ρ}(i{)}_{\mathrm{down}}$

$\stackrel{\_}{\mathrm{\ρ}}\left(d\right)=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\Σ}}}\mathrm{\ρ}\left(i\right)$

In the formula: ρ (d) _UpBe up navigation channel traffic density in certain day, ρ (d) _DownBe descending navigation channel traffic density in certain day, ρ (d) is a up-downgoing navigation channel traffic density in certain day;

Statistics monthly average navigation channel traffic density or annual navigation channel traffic density removed last corresponding fate then as long as the navigation channel traffic density of up-downgoing every day in certain month, certain year added up to add up.

Beneficial effect of the present invention mainly shows: 1, by the master data that many visual sensings merge and computing machine is collected traffic on inland waters automatically, can carry out effective traffic on inland waters management; 2, can be on macroscopic view and microcosmic understand and grasp actual state, essential characteristic and the universal law of traffic on inland waters, for the traffic control in inland river, traffic administration, traffic programme, waterway network construction, supervision of law enforcement etc. provide reference.

Description of drawings

Fig. 1 is the schematic diagram of the ship automatic identification system of multiple vision sensor information fusion;

Fig. 2 is the navigation channel Video Detection and the detection line configuration schematic diagram of the ship automatic identification system of multiple vision sensor information fusion;

Fig. 3 is the name of vessel sign trade mark in the boats and ships integral body and the synoptic diagram of draft;

Fig. 4 is the process flow diagram of name of vessel sign trade mark identifying;

Fig. 5 is the schematic diagram of the another kind of scheme of ship automatic identification system of multiple vision sensor information fusion;

Fig. 6 is the schematic diagram that the ship automatic identification system of multiple vision sensor information fusion adopts the scheme of omnibearing vision sensor;

Fig. 7 is the process flow diagram of the ship automatic identification system of multiple vision sensor information fusion;

Fig. 8 is BP neural network structure figure;

Fig. 9 is a BP neural network structure synoptic diagram;

Figure 10 is that non-maximum value suppresses the computing method synoptic diagram;

Figure 11 is a boats and ships object multiple target tracking processing flow chart;

Figure 12 is video image figure and the control area division synoptic diagram that omnibearing vision sensor obtains;

Figure 13 is an omnibearing vision sensor catadioptric schematic diagram;

Figure 14 is the structure principle chart of omnibearing vision sensor;

Figure 15 is the precedence diagram of wide-angle imaging machine calibration process conversion;

Figure 16 is the graph of a relation between the imaging system coordinate system of wide-angle imaging machine;

Figure 17 is the rotational transform synoptic diagram of picture plane in actual coordinates of wide-angle imaging machine.

Embodiment

Below in conjunction with accompanying drawing the present invention is further described.

Embodiment 1

With reference to Fig. 1～Figure 17, plant the inner river ship automatic identification system that multiple vision sensor information merges, comprise a plurality of vision sensors, be used for boats and ships to image tracing, the microprocessor of image recognition and traffic on inland waters situation statistics and be used to show the ship tracking situation, the boats and ships general image, ship-handling cabin image, the display unit of traffic on inland waters statistical graph etc., described a plurality of vision sensor connects microprocessor by video card, described a plurality of vision sensor comprises that one monitors large-range monitoring vision sensor 1 and a fire ball vision sensor 2 that carries out feature candid photograph hull image and driving cabin (name of vessel sign board) image in the navigation channel, described microprocessor comprises: image-display units is used for showing the video image of whole cruiseway, follow the tracks of the general image of boats and ships and the driving cabin image of following the tracks of boats and ships; Described large-range monitoring vision sensor 1 is an omnibearing vision sensor, and this omnibearing vision sensor is installed in the centre in navigation channel, is used for monitoring the tracking boats and ships in whole navigation channel; Described large-range monitoring vision sensor 1 is the wide-angle vision sensor, is installed in a side in navigation channel, is used for monitoring the tracking boats and ships in whole navigation channel; Described fire ball vision sensor 2, being used for that shippping traffic is carried out feature captures, obtain the locus at boats and ships place by the tracking of large-range monitoring vision sensor 1, microprocessor is captured then according to the direction rotation focusing of this positional information indication fire ball vision sensor 2 towards the locus at boats and ships place; Information fusion between described large-range monitoring vision sensor 1 and the fire ball vision sensor 2 realizes by mapping table; Described omnibearing vision sensor comprises the evagination catadioptric minute surface that is used for reflecting field, navigation channel object, evagination catadioptric minute surface down, be used to prevent anaclasis and the saturated dark circles cone of light, the dark circles cone is fixed on the center of catadioptric minute surface male part, be used to support the transparent cylinder of evagination catadioptric minute surface, be used to take the camera of imaging body on the evagination mirror surface, camera facing to the evagination mirror surface up; Described microprocessor also comprises: large-range monitoring vision sensor 1 demarcating module is used to set up the navigation channel chart picture in space and the corresponding relation of the video image that is obtained; Video data Fusion Module between large-range monitoring vision sensor 1 and the fire ball vision sensor 2 is used to control the rotation and the focusing of fire ball vision sensor 2, makes fire ball vision sensor 2 can aim at the tracking boats and ships and carries out the feature candid photograph; The multiple target tracking module is used to follow the tracks of the passing boats and ships on the navigation channel; The dummy line customized module, be used to customize the detection line on the monitoring navigation channel, customize three among the present invention and detected dummy line I, II and III, article three, detect that dummy line equates (actual range on the face of navigation channel) at interval and perpendicular to the navigation channel, article three, the detection dummy line all in the field range of large-range monitoring vision sensor 1, intersect by a wherein middle installation site of detecting dummy line and fire ball vision sensor 2; Boats and ships enter and detect the dummy line detection module, are used for system and produce an incident automatically, all can call corresponding processing module when system whenever produces an incident automatically; Boats and ships ID number and deposit the automatically-generating module of following the tracks of the boats and ships image folder, be used for the boats and ships of the dummy line of the visual field ragged edge that just enters large-range monitoring vision sensor 1 are named, and generate a file simultaneously with ID number name of these boats and ships, be used to deposit the close-up image of these boats and ships; Module is captured in boats and ships general image location, is used to locate the image of the integral body of capturing these boats and ships; The boats and ships profile detection module is used to detect the edge contour of these boats and ships and estimate the height that this boats and ships hull surfaces; Shipping draft and tonnage estimation block, be used to estimate the dead weight capacity of these boats and ships, the height that surfaces according to this boats and ships hull and leave data such as these boats and ships in the boats and ships master database are long, high, wide in, calculate this shipping draft, then the estimation dead weight capacity that calculates these boats and ships from the length and the width data of this ship; Module is captured in ship-handling cabin location, is used to locate the approximate location of the suspension name of vessel sign board of these boats and ships or lamp box and this position is captured the image of the name of vessel sign trade mark; Name of vessel sign trade mark identification module is used to discern the name of vessel sign trade mark of shippping traffic and obtains this ship basic information for the retrieval of follow-up master database from boats and ships and obtains retrieving major key; Boats and ships essential information retrieval module, being used to obtain with the name of vessel sign trade mark is the boats and ships essential information of major key; Shippping traffic record generation module is used for generating automatically a shippping traffic record, thereby is that statistics such as traffic on inland waters amount etc. performs the data preparation; Average volume of traffic statistical module is used to add up hourly traffic volume, daily traffic volume, the monthly average volume of traffic or the annual volume of traffic by certain point on the navigation channel; The deadweight tonnage statistical module is used to add up hour deadweight tonnage, day deadweight capacity position, monthly average deadweight tonnage or annual deadweight tonnage by certain point on the navigation channel; The traffic density statistical module is used to add up hour traffic density, daily traffic density, month traffic density or year traffic density by certain point on the navigation channel; Testing result affirmation, change, completion module are used to confirm that the identification of the name of vessel sign trade mark recognition result, the completion whether correct, the change identification error of shippping traffic do not have the name of vessel sign trade mark of identification or default boats and ships attribute master data;

Described boats and ships ID number, be used to produce one and can identify the major key that the shippping traffic object followed the tracks of, storage write down some other attribute data of the time that enters the monitoring field, course, speed and the ship of these boats and ships, boats and ships ID number naming rule is: YYYYMMDDHHMMSS* names with 14 bit signs, and YYYY-represents the year of the Gregorian calendar; MM-represents the moon; DD-represents day; HH-represents hour; MM-represents branch; SS-represents second; Automatically produce by the system for computer time;

The described automatically-generating module of following the tracks of the boats and ships image folder of depositing, be used to preserve the close-up image of the shippping traffic object of tracking, when the boats and ships object enters monitoring range (dummy line of outermost), system automatically produce one boats and ships ID number, in certain deposits the file of image, create simultaneously one with boats and ships ID number file of the same name, be used for depositing the close-up image of these boats and ships, the naming method of image file is: 1,3,5 general images of depositing the boats and ships object of being captured; 2,4,6 close-up image of depositing the driving cabin of the boats and ships object of being captured;

Module is captured in described boats and ships general image location, is used to locate the image of the integral body of capturing these boats and ships; We follow the tracks of every shippping traffic in described multiple target tracking module, when the boats and ships object of following the tracks of enters the detection dummy line, system can be according to the center of large-range monitoring vision sensor 1 detected boats and ships object, information such as size send instructions and carry out the candid photograph of boats and ships general image location to 2 pairs of boats and ships objects of fire ball vision sensor, the general image of capturing each boats and ships object among the present invention has 3, the general image of the boats and ships object of stravismus and positive side position about being respectively, the filename that the general image of the boats and ships object of being captured is deposited is respectively 1,3,5;

Described boats and ships profile detection module, be used to detect the shippping traffic object in the height of the water surface and the position of finding the driving cabin place, the border of hull is a very important descriptor of describing hull load-carrying feature and driving cabin feature, and these borders may produce marginal information in imaging process.The edge is meant the combination that those pixels of significant change are arranged in its surrounding pixel gray scale.The edge is the vector with amplitude and direction, and it shows as the sudden change of gray scale in image.Rim detection is exactly the noncontinuity that will detect this gray scale in the image.Therefore after after resume module is captured in described boats and ships general image location, obtaining the close-up image of hull,, the close-up image of this hull obtains hull load-carrying feature and driving cabin characteristic information thereby being carried out rim detection; Adopt optimum notch cuttype edge detection algorithm among the present invention. be the canny edge detection algorithm.The process step of Canny edge detection algorithm is as follows:

Step 1: use the Gaussian filter smoothed image

Adopt separable filtering method to ask the convolution of image and Gauss's smothing filtering, obtain the smoothed data array, meter bamboo mat formula is shown in (2):

G(x，y)=f(x，y)×H(x，y) (2)

In the formula: wherein beat (x y) is Gauss's smothing filtering function, $H(x,y)=e^{-\frac{a^2+{\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">b</figure-callout>}}^2}{{2\mathrm{\&delta;}}^2}}$

(x y) is original image to f, and (x y) is the result who carries out convolution to G.

Step 2: with the finite difference of single order local derviation assign to the amplitude of compute gradient and direction to gaussian filtering after each pixel in the image, the compute gradient amplitude

And direction

Computing formula as the formula (3),

H wherein ₁, H ₂Be respectively level and VG (vertical gradient) operator, Carry out the result of convolution with original image;

Step 3: gradient magnitude is carried out non-maximum value suppress;

It is the gradient of the overall situation that above-mentioned calculating only obtains, and is not sufficient to determine the edge.In order to determine the edge of boats and ships, must keep the point of partial gradient maximum, and suppress non-maximum value.Adopt among the present invention gradient magnitude is carried out " non-maximum value inhibition ".The label of four sectors shown in the accompanying drawing 18 is 0 to 3, and therefore four kinds of combinations of corresponding 3*3 neighborhood are arranged.

To a point, the center pixel M of neighborhood compares with two pixels along gradient line.If the Grad of M is big unlike two adjacent image point Grad along gradient line, then make M=O.

Step 4: use the detection of dual threshold algorithm and be connected the edge;

In order to reduce the quantity of false edge section, adopt the dual threshold algorithm, all values that will be lower than threshold value is composed null value.The dual threshold algorithm suppresses two value of cutting off from τ 1 of image effect and τ 2 to non-maximum value, and 2 τ, 1 ≈ τ 2, thereby can obtain two threshold value edge image N1[i, i] and N2[i, j].Since N2[i, j] use high threshold to obtain, thereby contain false edge seldom, but interruption (not closed) is arranged.The dual threshold method will be at N2[i, j] in the edge is connected into profile, when arriving the end points of profile, this algorithm is just at N1[i, j] 8 adjoint point positions seek the edge can be connected on the profile, like this, algorithm is constantly at N1[i, j] in collect the edge, up to N2[i, j] couple together till.

Described multiple target tracking module is used to follow the tracks of the passing boats and ships on the navigation channel, for the headway of Ship ', the detection of variety of event and the location candid photograph of boats and ships provide technical foundation; Background is the navigation channel among the present invention, and tracking target is to liking boats and ships.Therefore at first to extract the pixel of prospect boats and ships in the navigation channel background at the low-level image feature layer.Extract the method for the pixel of boats and ships and mainly be made up of white adaptation background subtraction, shade inhibition and three parts of connected region sign, its precedence diagram as shown in Figure 11;

Described adaptive background is subdued, be used for cutting apart in real time dynamic object (underway boats and ships), the adaptive background elimination algorithm that is based on mixture gaussian modelling that we adopt, its basic thought is: the feature of using mixture gaussian modelling to come each pixel in the phenogram picture frame; When obtaining new picture frame, upgrade mixture gaussian modelling; On each time period, select the subclass of mixture gaussian modelling to characterize current background; If the pixel of present image and mixture gaussian modelling are complementary, judge that then this point is a background dot, otherwise judge that this point is the foreground point.

Detect at the brightness value Y component in the YCrCb color space of image.The ADAPTIVE MIXED Gauss model has adopted the hybrid representation of a plurality of Gauss models to each picture point, establishes total K of the Gaussian distribution that is used for describing each some color distribution, is labeled as respectively

η(Y _i，u _t，i，Σ _t，i)，i＝1,2，3…，k (4)

Subscript t express time in the formula (4).Each Gaussian distribution has different weights and priority respectively, K background model is sorted according to priority order from high to low again, gets suitable surely background model weights and threshold value.When detecting the foreground point, according to priority order with Y _iMate one by one with each Gaussian distribution model.If coupling is judged that then this point may be the foreground point, otherwise is the foreground point.If certain Gaussian distribution and Y _tCoupling is then upgraded by certain turnover rate weights and Gauss's parameter of this Gaussian distribution.

Described connected region sign is used to extract prospect boats and ships object, and the method for connected region sign is a lot, and what adopt among the present invention is eight connected region extraction algorithms.The connected region sign is subjected to the noise effect in the primary data very big in addition, generally need carry out denoising earlier, denoising can realize by morphology operations, utilize corrosion and expansion operator to remove isolated noise foreground point and the aperture of filling up the target area respectively herein, specific practice is: offset earlier except the prospect point set F after the background model and expand respectively and corrosion treatment, obtain expansion collection Fe and shrink collecting Fc, by handling resulting expansion collection Fe and shrinking the result that collection Fc can think initial prospect point set F is filled up aperture and removal isolated noise point.Therefore there is the following Fc of relation＜F＜Fe to set up, then to shrink collection Fc as starting point, on expansion collection Fe, detect connected region, then testing result is designated as { Rei, i=1,2,3,, n}, the connected region that will detect gained at last projects on the initial prospect point set F again, get communication with detection result { Ri=Rei ∩ F to the end, i=1,2,3, n} can keep the integrality of target also to avoid the influence of noise foreground point simultaneously by this target partitioning algorithm, has also kept the edge details part of target.

Just can from the navigation channel background of static state, dynamic prospect boats and ships object extraction be come out through above three steps, and can obtain some initial informations, as the size of boats and ships object and position or the like based on the boats and ships object.Extract prospect boats and ships object in the background of navigation channel after, the processing of following is that the boats and ships object in the navigation channel is followed the tracks of processing; Adopt based target color characteristic track algorithm among the present invention, this algorithm is the further improvement to the MEANSHIFT algorithm, in this algorithm, utilize the color characteristic of ship target object in video image, to find the position and the size at motion ship target object place, in the next frame video image, with moving target current position and big or small initialization search window, repeat this process and just can realize Continuous Tracking target.The initial value of search window is set to current position of moving target and size before each the search, because search window is searched near just marking the zone that may occur in motion day, so just can save a large amount of search times, make this algorithm have good real time performance.This algorithm is to find moving target by color-match simultaneously, and in the process of moving target motion, colouring information changes little, so this algorithm has good robustness.The orientation of resulting ship target object and size after the connected region identification process are submitted to result color of object signature tracking algorithm then, with realize the boats and ships object from motion tracking, as shown in Figure 11.

Module is captured in described ship-handling cabin location, is used to locate the approximate location of the suspension name of vessel sign board of these boats and ships or lamp box and this position is captured the image of the name of vessel sign trade mark; The 4th, the 5th regulation according to " spitkit name of vessel flag management tentative method ", spitkit should be aboard ship significantly the position indicate and its registration name of vessel, the consistent name of vessel and the port of registry, the port of registry, and accept the supervision and check of maritime affairs at different levels management organization; Regulation indicates that name of vessel should indicate name of vessel at bow two sides of a ship and stern in inland navigation craft, perhaps be installed in the suitably remarkable position of pilothouse superstructure top or left and right sides bulkhead towards side of a ship external symmetry, perhaps face forward is installed suitably remarkable position aboard ship, perhaps be equipped with the appropriate location that lamp box surrogate markers board is installed in axis, pilothouse superstructure top, to discern the name of vessel of all shippping traffics, need obtain the image of the various remarkable positions of boats and ships from different perspectives, could be from bow two sides of a ship, stern, the superstructure top of pilothouse or left and right sides bulkhead, obtain image information on the appropriate location of remarkable position, the place ahead of boats and ships and axis, pilothouse superstructure top about name of vessel; The position that can see name of vessel from this regulation is big or small closely-related with boats and ships, if the captain who detects just captures the fire ball camera head more than 20 meters facing to bow two sides of a ship and stern; If the boats and ships of the captain who detects below 20 meters at first will detect these boats and ships by edge detection algorithm whether the cabin is arranged, for the boats and ships that the cabin is arranged, the name of vessel sign board may be installed in the suitably remarkable position or the face forward of pilothouse superstructure top or left and right sides bulkhead suitably remarkable aboard ship position is installed, also can be installed in the appropriate location of axis, pilothouse superstructure top in addition with lamp box surrogate markers board, pilothouse and two kinds of situations of back pilothouse before the pilothouse of the boats and ships that navigate by water in the actual inland river also exists simultaneously; Therefore the position that how to detect pilothouse exactly is then to the location of name of vessel sign board, discrimination to the raising name of vessel sign trade mark is very favorable, according to the 11 regulation in " spitkit name of vessel flag management tentative method ", the name of vessel sign board uses white edge wrongly written or mispronounced character of the blue end, and front background color and word look should be reflective; Article 12, regulation, name of vessel lamp box side display board is white background The Scarlet Letter, the inner daylight lamp of installing, therefore utilize the color characteristic of name of vessel sign board and name of vessel lamp box to realize name of vessel sign board and the candid photograph of name of vessel lamp box location, discrimination to the raising name of vessel sign trade mark also is very favorable, the image of capturing name of vessel sign board and name of vessel lamp box for every shippping traffic in this patent has three at least, wherein an image is positive side position, other two images be respectively about stravismus position, about the stravismus bit image be mainly used in face forward be installed in the suitably significantly name of vessel sign board of position identification on the ship; Capture location, ship-handling cabin is to be based upon on effective tracking basis of ship target object, when tracked ship target object process detected dummy line, computing machine sent a series of instruction control fire ball vision sensors 2 automatically and turns to the residing orientation of ship target object, good focal length is captured then; Candid photograph to the ship target object in the present invention has two kinds of different purposes, and the first is used to estimate the dead weight capacity of boats and ships object, and it two is the identification that is used for the name of vessel sign trade mark; Therefore the former need capture the general image of boats and ships object, and the latter need capture the location drawing picture at the driving cabin place of boats and ships object; At first be after capturing the general image of boats and ships object in the present invention, according to the color characteristic of driving cabin contour feature on the general image of boats and ships object and name of vessel sign board the driving cabin of boats and ships object is located the back fast then and capture the name of vessel sign board, handle so that carry out the identification of the name of vessel sign trade mark; The ship-handling cabin image of capturing each boats and ships object among the present invention has 3, be respectively about the ship-handling cabin image of boats and ships object of stravismus and positive side position, the filename that the ship-handling cabin image of the boats and ships object of being captured is deposited is respectively 2,4,6;

Described name of vessel sign trade mark identification module is used to discern the name of vessel sign trade mark of shippping traffic and obtains this ship basic information for the retrieval of follow-up master database from boats and ships and obtains retrieving major key; In name of vessel sign trade mark identification module, comprise the positioning unit based on the name of vessel sign board of color model, the extraction unit of name of vessel sign board, the correcting unit of name of vessel sign board, the Character segmentation unit and the character recognition unit of name of vessel sign board;

According to the tenth regulation in " spitkit name of vessel flag management tentative method ", name of vessel sign board, the employed Chinese character of lamp box are the standard simplified Chinese character, and font is upright director circle body, English alphabet and arabic numeral for etc. bold face letter.According to name of vessel sign board of the 9th regulation, lamp box content is the name of vessel and the port of registry, writes arrangement placed in the middle, and name of vessel is last, and the port of registry is down; What be concerned about among the present invention is obtaining of name of vessel, and name of vessel is divided into two hurdles in the sign board name of going on board, above a hurdle administrative region, expression boats and ships place and the description of ship sign formed by Chinese character, below the number formed by English alphabet and arabic numeral of a hurdle; Because Chinese character adopted in sign board font is different with the font that English alphabet and arabic numeral are adopted, in order to improve discrimination, in this patent, Chinese Character Recognition and English alphabet and arabic numeral is discerned to make a distinction and handle; We adopt the K-NN nearest neighbor classifier in Chinese Character Recognition, and the structure of the standard Chinese character of upright director circle body is made a standard form class, the standard Chinese character in Chinese character to be identified and the standard form class is compared realize the identification of Chinese character;

The positioning unit of described name of vessel sign board based on color model is used for determining the position of name of vessel sign board at hull; According to the regulation of marine board of the People's Republic of China (PRC), the name of vessel sign board uses white edge wrongly written or mispronounced character of the blue end, and the side display board of name of vessel lamp box is white background The Scarlet Letter; Adopt LUV color model and edge contour algorithm to come station keeping ship name sign board among the present invention based on even color space; Whether slightly judge at first earlier has blueness or white to exist, if exist comparatively significantly blue portion, jointing edge profile algorithm is determined the position of name of vessel sign board again in the image of being captured; If exist comparatively significantly white portion, jointing edge profile algorithm is determined the position of name of vessel sign board again, the result of location as Fig. 4 a) shown in;

The extraction unit of described name of vessel sign board is used for the name of vessel sign board is extracted separately from hull; No matter the name of vessel sign board still is the name of vessel lamp box, and their color is complementary, and blue background is corresponding to wrongly written or mispronounced character, the background of white is corresponding to The Scarlet Letter, all exist complementary relationship, utilize its complementary relationship to extract the name of vessel sign board among the present invention, the result of extraction is as the b of Fig. 4) shown in;

The correcting unit of described name of vessel sign board is used for that institute is extracted the rectification of name of vessel sign board image and is the front elevation picture; When obtaining the name of vessel sign board because the angle of capturing is indefinite, captured name of vessel sign board image can exist angle of inclination and deformation angle, by the hough conversion rectangular coordinate is transformed into the parameter coordinate system among the present invention, utilize a kind of new K average fitting a straight line algorithm, detect these two angles of ship board; And then the method for utilization " rotation+mistake is cut " conversion, promptly elder generation is rotated the angle of inclination and does to correct for the first time, carries out wrong shearization again and does the rectification second time.At last deformation angle is carried out stretching conversion, whole licence plate is corrected into the licence plate of facing of a standard, the result of rectification is as the c of Fig. 4) shown in;

The Character segmentation unit of described name of vessel sign board is used for the character on the name of vessel sign board after correcting is cut apart; A kind of coupling of masterplate is fast cut apart each character on the name of vessel sign board in conjunction with the vertical and horizontal sciagraphy among the present invention.Owing to proofread and correct the front elevation picture produced standard, literal big or small similar, masterplate mates can accurately cut apart literal.The speed of masterplate matching algorithm can be improved in conjunction with the vertical and horizontal projection.Because the inevitable gap location in intercharacter or character of the projection of character block in the vertical direction is obtained local minimum, therefore, the correct split position of character should be near above-mentioned local minimum, and this position should be satisfied in [2006] No. 594 files of seagoing vessel oceangoing ship of marine board of the People's Republic of China (PRC) about character format write, character size restriction and some other conditions of name of vessel sign board, and the result of cutting apart is as the d of Fig. 4), e) shown in;

Described character recognition unit is used for the character of each name of vessel sign board after cutting apart is discerned; Before identification, at first to carry out the character normalized, character is normalized to unified size, discern the different recognition technology of employing in this patent for Chinese character and English alphabet, arabic numeral, for the recognition technology that has adopted sorter in the Chinese Character Recognition this patent; For the recognition technology that has adopted in English alphabet, the arabic numeral identification this patent based on the BP neural network; The recognition technology of described sorter is the Chinese character after the normalized and standard Chinese character to be mated discern;

Described standard Chinese character is that (captain is more than 20 meters: height: 110, wide: 90, stroke is thick: 11 with upright director circle body, font size according to the Chinese character that is occurred in the name of vessel sign board (type name of China's administrative area domain name+ship); The captain is below 20 meters: height: 92, wide: 75, stroke is thick: 9; No cabin canoe: height: 55, wide: 45, stroke is thick: 7) customize, these standard Chinese characters that customize are left in be used in the standard Chinese character storehouse mating during Chinese character in identification;

Described sorter adopts the K-NN nearest neighbor classifier, promptly for the sample of a classification the unknown, can suppose that its classification is the classification at the nearest training sample of this sample of feature space middle distance, and in most of the cases, this hypothesis is rational.Under the more situation of classification, difference varies between different classes of, if remove the testing classification device with different master samples, class for correct identification, should produce identical result, promptly the simplest sample class and sample with complex class, sorter will produce same output to the master sample that can correctly discern.Suppose to have n class standard sample class, we can define a _n{ n=1,2,3 ... n} is the complexity a of master sample class _MaxGet a respectively _Max=Max (a _n), n=1,2,3 ... n, a _Min=Min (a _n), n=1,2,3 ... n makes a _MinBeing the weights between inhomogeneity, so just making sorter have identical output for the master sample that can correctly discern, promptly is the normalization operation.

And for the Chinese Character Recognition under some non-idealities, the structure of the font of testing is the match-on criterion class fully, therefore can have two kinds of situations.Promptly with go the to make a strategic decision simple sample and go the complex samples of making a strategic decision with simple master sample class of complicated master sample class, both of these case all can produce very big risk.But no matter for which kind of situation, should not obtain correct output, should produce the difference that differs greatly with correct output in other words, this just need do punishment to sorter and judge.

For the master sample class of complexity, less relatively for the punishment of latter event, and should strengthen relatively for the punishment of the previous case, and for simple master sample class, then just opposite.Order

Be the punishment to latter event, c _nN=1,2,3 ... n} is the punishment to the previous case, a _n, c _n=k ^*a _n, k is a scale-up factor.The purpose of so doing is in order to make all kinds of difference big as much as possible, to make that the effect of classification is more obvious.

By above sorter the Chinese character that is occurred in the sign board is discerned, wherein the standard sample given figure of sorter depends on the number for the Chinese character that is occurred, and can also further determine weights by the characteristic of output image simultaneously, to reach the better recognition effect.

Described English alphabet and arabic numeral identification, be used for being identified in English alphabet and the arabic numeral that name of vessel uses, compare with Chinese Character Recognition, English alphabet has only 26, numeral has only 10, and stroke is also simple, realizes relatively easily for identification, and sample number has only 36, belongs to the identification of small sample; Because the BP neural network has the discrimination height to small sample, adopt the BP neural network that English alphabet and arabic numeral are discerned in this patent;

Described BP neural network is used for discerning English alphabet and the arabic numeral that name of vessel uses, and BP neural network structure figure as shown in Figure 9; In the ordinary course of things, the BP neural network is made of an input layer, one or more hiding layer, an output layer etc., as shown in Figure 8;

1) input/output relation

I-〉H (input layer is to hiding layer)

a ₂＝purelin(LW _2，1，a ₁+b ₂)a ₁＝tansig(IW _1，1，p ₁+b ₁)

tan sig(x)＝tanh(x)＝(e ^x-e ^-x)/(e ^x+e ^-x)

H-〉0 (hide layer and lead output layer)

a ₂＝purelin(LW _2，1，a ₁+b ₂)

Wherein, defeated people's layer neuron number is n, and the hidden layer neuron number is n1, and the output layer neuron number is S2.

2) network learning and training

One of key issue of BP neural network is the definite of weights, and the weights in the BP neural network determine that method is as follows,

At first supposition input q organizes this P ₁, P ₂..., P _q, P _i∈ R _n,

Desired output is T ₁, T ₂..., T _q, T ∈ Rs here ₂,

The actual a that is output as of network ₂₁, a ₂₂..., a _2q, a ₂∈ Rs ₂

The criterion of estimating is the error minimum, so the essence of network learning and training is converted into an optimization problem.Here the relation that consideration comes error identifying and weighting coefficient with gradient method (Gradient) is to obtain the rule that weighting coefficient changes.

The definition error function is:

$E(W,B)=\frac{1}{2}\underset{k=1}{\overset{s2}{\mathrm{\&Sigma;}}}(t_k-{a2}_k{)}^2$

Utilize the tonsure descent method to ask the variation of weights and the backpropagation of error then.

(1). the weights of output layer change from i weights change that is input to k output to be had:

${\mathrm{\&Delta;<figure-callout\; id="2"\; label="w"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">w</figure-callout>}2}_{\mathrm{ki}}=-\mathrm{\&eta;}\frac{\&PartialD;E}{{\&PartialD;\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">w</figure-callout>}2}_{\mathrm{ki}}}=-\mathrm{\&eta;}\frac{\&PartialD;E}{{\&PartialD;a2}_K}\frac{{\&PartialD;a2}_k}{{\&PartialD;\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">w</figure-callout>}2}_{\mathrm{ki}}}$

$=\mathrm{\&eta;}(t_k-{a2}_k)\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">f</figure-callout>}2\&prime;{a1}_i$

$={\mathrm{\&eta;\&delta;}}_{\mathrm{ki}}{a1}_i$

In like manner can get:

${\mathrm{\&Delta;<figure-callout\; id="2"\; label="b"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">b</figure-callout>}2}_{\mathrm{ki}}=-\mathrm{\&eta;}\frac{\&PartialD;E}{{\&PartialD;\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">b</figure-callout>}2}_{\mathrm{ki}}}=-\mathrm{\&eta;}\frac{\&PartialD;E}{{\&PartialD;a2}_k}\frac{{\&PartialD;a2}_k}{{\&PartialD;\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">b</figure-callout>}2}_{\mathrm{ki}}}$

$=\mathrm{\&eta;}(t_k-{a2}_k)\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">f</figure-callout>}2\&prime;$

$={\mathrm{\&eta;\&delta;}}_{\mathrm{ki}}$

(2). the hidden layer weights change

${\mathrm{\&Delta;<figure-callout\; id="1"\; label="w"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">w</figure-callout>}1}_{\mathrm{ij}}=-\mathrm{\&eta;}\frac{\&PartialD;E}{{\&PartialD;\mathrm{<figure-callout\; id="1"\; label="W"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">W</figure-callout>}1}_{\mathrm{ij}}}=-\mathrm{\&eta;}\frac{\&PartialD;E}{{\&PartialD;a2}_k}\frac{{\&PartialD;a2}_k}{{\&PartialD;a1}_j}\frac{{\&PartialD;a1}_i}{{\&PartialD;\mathrm{<figure-callout\; id="1"\; label="w"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">w</figure-callout>}1}_{\mathrm{ij}}}$

$=\mathrm{\&eta;}\underset{k=1}{\overset{s2}{\mathrm{\&Sigma;}}}(t_k-{a2}_k)\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">f</figure-callout>}2\&prime;{\mathrm{<figure-callout\; id="2"\; label="w"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">w</figure-callout>}2}_{\mathrm{<figure-callout\; id="1"\; label="ki\; f"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">ki</figure-callout>}}f1\&prime;p_j$

$={\mathrm{\&eta;\&delta;}}_{\mathrm{ij}}{p}_j$

Wherein, δ _Ij=e _iF1 ', $e_i=\underset{k=1}{\overset{\mathrm{<figure-callout\; id="2"\; label="s"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">s</figure-callout>}2}{\mathrm{\&Sigma;}}}{\mathrm{\&delta;}}_{\mathrm{<figure-callout\; id="2"\; label="ki\; w"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">ki</figure-callout>}}{w}_{}{2}_{\mathrm{ki}}$

In like manner can get Δ bl _i=η δ _Ii

(3). explain

The output layer error e _j(j=1-S ₂)

The hidden layer error e _i(i=1-n ₂),

Through above-mentioned processing, the weights of error minimum are exactly the weights of expectation;

Described boats and ships essential information retrieval module, being used to obtain with the name of vessel sign trade mark is the boats and ships essential information of major key; In the boats and ships Basic Information Table, include: shipRegisterCode: Official Number, chineseShipName: Chinese name of vessel, englishShipName: English name of vessel, originalChineseName: former Chinese name of vessel, shipCallLetters: ship'call sign, the IMOCode:IMO numbering, shippingAreaCode: navigating area code, shipNativePortCode: boat nationality port code, originalShipNativePort: former boat nationality port, originalShipRegisterCode: former Official Number, firstRegisterCode: first registration number, shipTypeCode: Ship Types code, hullMaterialCode: hull material code, buildingPlace: shipbuilding place, buildingPlaceEnglish: English shipbuilding place, rebuildPlace: reconstruction place, rebuildPlaceEnglish: English reconstruction place, rebuildDate: reconstruction date, dockyard: shipyard, dockyardEnglishName: English shipyard, buildedDate: creation data, shipLength: boats and ships length, shipWidth: boats and ships width, shipHigh: the boats and ships degree of depth, totalTon: gross ton, netTon: deadweight ton, TotalLoadTon: dead weight all hold, mainFrameTypeCode: main frame kind category code, mainFrameNum: host number, mainFramPower: main engine power, thrusterTypeCode: information such as thruster kind category code, the name of vessel sign trade mark that uses among the present invention is corresponding to the chineseShipName in the boats and ships master data table: Chinese name of vessel can obtain shipLength by retrieval: boats and ships length, shipWidth: boats and ships width, shipHigh: the boats and ships degree of depth, totalTon: gross ton, netTon: deadweight ton, TotalLoadTon: information such as dead weight all hold; These information are used for the draft and the dead weight capacity of follow-up boats and ships and judge;

Described shipping draft and tonnage estimation block, be used to estimate the dead weight capacity of these boats and ships, the height that surfaces according to this boats and ships hull and leave data such as these boats and ships in the boats and ships master database are long, high, wide in, extrapolate this shipping draft, then the estimation dead weight capacity that calculates these boats and ships from the length and the width data of this ship; Estimation equation is: ship side is exposed at the height on the water surface during draft during Ship's Cargo draft=boats and ships ship side overall height-boats and ships hole capital after selling all securities-boats and ships loading; Water discharge (the metric ton)=wide * goods of long * draft * side's mode coefficient (cubic meter)/0.9756 (seawater) or 1 (fresh water) (cubic meter), displacement tonnage is as the deadweight tonnage of estimation boats and ships;

Described shippping traffic record generation module is used for generating a shippping traffic record automatically, prepares thereby traffic on inland waters amount statistics etc. perform data; The shippping traffic recorded content is as shown in the table:

Table 1 shippping traffic record sheet

Field name	Data type	Explanation
			Boats and ships ID	String	Major key
The name of vessel sign trade mark	String
			The captain	Float
The beam	Float
			Height	Float
The estimation tonnage	Float
			Detection time	Data
Speed	Float
			On/descending	Boolean

In the table 1 boats and ships ID be boats and ships when entering in the monitoring field system produce automatically; The name of vessel sign trade mark obtains after name of vessel sign trade mark identification module is handled; Captain, the beam, height are to retrieve resulting by the name of vessel sign trade mark in boats and ships essential information retrieval module; The estimation tonnage be in shipping draft and the tonnage estimation block by shipping draft, the estimation that calculates these boats and ships from the length and the width data of this ship is resulting then; Be that to enter in the dummy line detection module auto acquisition system time at boats and ships resulting detection time; Speed calculates by the speed of the ship in metres per second detecting unit; On/descendingly calculate by boats and ships up-downgoing judging unit;

Described speed of the ship in metres per second detecting unit is used to detect the headway of a certain boats and ships, when boats and ships respectively when detecting dummy line I, II and III, we can obtain a complete boats and ships object information: Ship (boats and ships ID; ShipID, initial tracing positional; StartPosition, initial time; StartTime finishes tracing positional; EndPosition, the concluding time; EndTime), boats and ships ID enters at boats and ships and generates automatically when one of outermost detects dummy line, initial tracing positional is to enter the position (detect dummy line I or detect dummy line III) that one of outermost detects dummy line, initial time is to enter system time that detects dummy line of outermost, finishing tracing positional is to leave the monitoring field just to have run into the position (detect dummy line III or detect dummy line I) that one of outermost detects dummy line, concluding time is to leave the monitoring field just to have run into time of detecting dummy line of outermost, therefore has following formula to come the headway of Ship ':

$V\left(\mathrm{ShipID}\right)=\frac{\mathrm{EndPosition}-\mathrm{StartPosition}}{\mathrm{EndTime}-\mathrm{StartTime}}---\left(5\right)$

Described boats and ships up-downgoing judging unit, be used to judge the navigation direction of a certain boats and ships, if the boats and ships object of following the tracks of is at first run into is to detect dummy line I we are defined as up direction, is to detect dummy line III we are defined as down direction if the boats and ships object of following the tracks of at first runs into;

Further, Video Detection for waters, wide navigation channel, in order to obtain more video information, large-range monitoring vision sensor 1 among the present invention adopts omnibearing vision sensor, because omnibearing vision sensor can capture on the horizontal direction 360 ° video image, as shown in Figure 6, and the angle range of the horizontal direction of fire ball vision sensor 2 is 360 °, the angle range of vertical direction is 90 °, omnibearing vision sensor and fire ball vision sensor are merged, after omnibearing vision sensor is found certain ship motion subject object, system returns the coordinate information of a ship motion target, and utilize this information to control the fire ball vision sensor and take the ship motion target with the location, capture the detail image information of ship target and handle for follow-up identification.It is worthy of note, need to set in advance a spatial correspondence calibration scale, the fire ball vision sensor is located fast.Scaling method is: with the panoramic picture center is the center of circle, according to differentiating needs panoramic picture is divided into some annulus, as shown in figure 12, then each annulus is divided several equal portions, can divide according to actual needs.Like this, a width of cloth panoramic picture just has been divided into several zones regularly, and there are its specific angle, direction and size in each zone.Then, according to the parameters such as height, fire ball vision sensor locus and camera focal length of omnibearing vision sensor, determine regional required level, vertical angle and the focal length that rotates that the fire ball vision sensor will detect successively apart from the navigation channel water surface.Exist if omnibearing vision sensor has detected moving target, system just can regulate the fire ball vision sensor automatically according to this target region and calibration scale, and the ship motion target is captured feature; Can realize 360 ° real-time monitoring of horizontal direction in omnibearing vision sensor, its core component is the catadioptric minute surface, shown in 3 among Figure 13; Its principle of work is: enter the light at the center of hyperbolic mirror, reflect towards its virtual focus according to bi-curved minute surface characteristic.Material picture reflexes to imaging in the collector lens through hyperbolic mirror, a some P on this imaging plane (x, y) corresponding the coordinate A of a point spatially in kind (X, Y, Z).

3-hyperbolic curve face mirror among Figure 13,4-incident ray, the focus Om (0 of 7-hyperbolic mirror, 0, c), the virtual focus of 8-hyperbolic mirror is camera center O c (0,0 ,-c), 9-catadioptric light, 10-imaging plane, the volume coordinate A of 11-material picture (X, Y, Z), 5-incide the volume coordinate of the image on the hyperboloid minute surface, 6-be reflected in some P on the imaging plane (x, y).

Further, described catadioptric minute surface is in order to access the corresponding point with the space object coordinate, and the catadioptric minute surface adopts hyperbolic mirror to design: shown in the optical system that constitutes of hyperbolic mirror can represent by following 5 equatioies;

((X ²+Y ²)/a ²)-(Z ²/b ²)＝-1 (Z>0) (6)

$c=\sqrt{{\mathrm{\&alpha;}}^2+b^2}---\left(7\right)$

β＝tan ^-1(Y/X) (8)

α＝tan ^-1[(b ²+c ²)sinγ-2bc]/(b ²+c ²)cosγ (9)

$\mathrm{\&gamma;}={\tan }^{-1}\&lsqb;f/\sqrt{({\mathrm{<figure-callout\; id="2"\; label="x"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">x</figure-callout>}}^2+Y^2)}]$

In the formula, X, Y, Z representation space coordinate, c represents the focus of hyperbolic mirror, 2c represents two distances between the focus, a, b are respectively the real axis of hyperbolic mirror and the length of the imaginary axis, and β represents angle one position angle of incident ray on the XY plane, α represents angle one angle of depression of incident ray on the XZ plane, and f represents the distance of imaging plane to the virtual focus of hyperbolic mirror;

By formula (11), (12) can ((x, y), the information that can the implementation space and the conversion of imaging plane information be demarcated the purpose of the information of imaging plane to reach spatial information Z) to be mapped to coordinate points on the imaging plane for X, Y with the point on the space.

$X=\frac{{\mathrm{<figure-callout\; id="2"\; label="Xfa"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">Xfa</figure-callout>}}^2}{2\mathrm{bc}\sqrt{x^2+{\mathrm{<figure-callout\; id="2"\; label="Y"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">Y</figure-callout>}}^2+Z^2}-({\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">b</figure-callout>}}^2+c^2)Z}---\left(11\right)$

$y=\frac{{\mathrm{<figure-callout\; id="2"\; label="Yfa"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">Yfa</figure-callout>}}^2}{2\mathrm{<figure-callout\; id="2"\; label="bc\; X"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">bc</figure-callout>}\sqrt{X^{}}\sqrt{{}^2+{\mathrm{<figure-callout\; id="2"\; label="Y"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">Y</figure-callout>}}^2+Z^2}-({\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">b</figure-callout>}}^2+c^2)Z}---\left(12\right)$

Further, the fusion of multiple vision sensor information realizes by mapping table, the fire ball video camera of Xiao Shouing can be with 80-256 preset point in the market, apace the fire ball video camera is turned to detected target of omnibearing vision sensor and focusing rapidly in order to reach; In order to realize the fusion of multiple vision sensor information, visual field with omnibearing vision sensor in this patent is divided into 128 zones, as shown in Figure 12, each zone corresponding a preset point of fire ball video camera, therefore as long as on the captured video image of omnibearing vision sensor, divide 128 zones, then to the preset point of each regional center as the fire ball video camera; Detect boats and ships when certain surveyed area occurs at omnibearing vision sensor, control fire ball video camera arrived and this area relative preset point after microprocessor obtained this information, and the feature of boats and ships object is captured.

Described testing result affirmation, change, completion module, be used to confirm whether the name of vessel sign trade mark identification of shippping traffic correct, the recognition result of change identification error, completion some not have the name of vessel sign trade mark discerned or default boats and ships attribute master data; The identification of the name of vessel sign trade mark is because a variety of causes is not to come out by computer Recognition, also may exist some default data in the same boats and ships attribute master data, the means that therefore need to adopt manual intervention are imported these unidentified name of vessel sign trades mark that come out and default boats and ships attribute master data record; Designed an inputting interface among the present invention, the boats and ships image that shows 6 candid photographs in the middle and upper part, interface, wherein 3 be about the general image (file by name 1 of boats and ships object of stravismus and positive side position, 3,5), other 3 be about the cabin image (file by name 2 of boats and ships object of stravismus and positive side position, 4,6), the user can roughly determine the ship type according to the general image of about 3 stravismus and the boats and ships object of positive side position, table 2 is canal typical case's ship shape parameter, the user can select the ship type on one side with reference to the general image of boats and ships object on one side, behind the ship type that chooses these boats and ships, computing machine automatically is presented at these ship shape parameters on the interface, if shown parameter and reality do not match, the user also can revise these parameters;

Table 2 canal typical case's ship shape parameter

Ship type (t)	Freighter	Barge	Fleet
					The wide * drinking water of long * (m)	The wide * drinking water of long * (m)	The wide * drinking water of long * (m)	The formation type
	100	25*4.8*1.4	/	/					/
200					30*6.0*1.8	/	/	/
	300	36*7.2*1.8	33*6.8*1.8	417*6.8*1.8					One drags 12 to refute
400					42*7.4*2.0	38*7.3*1.8	527*7.3*1.8	One drags 12 to refute
	500	46*8.6*2.0	40*7.3*2.0	486*7.3*2.0					One drags 10 to refute
800					53*9.8*2.8	/	/	/
	1000	57*9.8*2.8	/	/					/

Further, the user can be according to 3 about the cabin image (file by name 2,4,6) of boats and ships object of stravismus and positive side position, the name of vessel trade mark that input shows on image in the input frame of the name of vessel trade mark or the incorrect name of vessel trade mark of modification identification, system calls boats and ships essential information retrieval module automatically behind the input name of vessel trade mark, thereby obtains the basic attribute data of these boats and ships;

Described volume of traffic statistical module is used to add up hourly traffic volume, daily traffic volume, the monthly average volume of traffic or the annual volume of traffic by certain point on the navigation channel; The volume of traffic is meant the boats and ships number that passes through a certain waters in the unit interval, and the size of the volume of traffic directly reflects the busy extent of this waters vessel traffic; Because the volume of traffic is time dependent, so generally adopt hourly traffic volume, daily traffic volume, the monthly average volume of traffic or the annual volume of traffic to represent.Hour the volume of traffic can use formula (13) expression,

$\stackrel{\_}{Q}(i{)}_{\mathrm{up}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{Q}_{\mathrm{up}}$

$\stackrel{\_}{Q}(i{)}_{\mathrm{down}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{Q}_{\mathrm{down}}$

$\stackrel{\_}{Q}\left(i\right)=\stackrel{\_}{Q}(i{)}_{\mathrm{up}}+\stackrel{\_}{Q}(i{)}_{\mathrm{down}}$

In the formula: i is the integer between 0～23,, Q (i) _UpBe up boats and ships number in certain i hour, Q (i) _DownBe descending boats and ships number in certain i hour, Q (i) is the boats and ships number of up-downgoing in certain i hour;

Daily traffic volume can be used formula (14) expression,

$\stackrel{\_}{Q}(d{)}_{\mathrm{up}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}Q(i{)}_{\mathrm{up}}$

$\stackrel{\_}{Q}(d{)}_{\mathrm{down}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}Q(i{)}_{\mathrm{down}}$

$\stackrel{\_}{Q}\left(d\right)=\stackrel{\_}{Q}(d{)}_{\mathrm{up}}+\stackrel{\_}{Q}(d{)}_{\mathrm{down}}$

In the formula: Q (d) _UpBe up boats and ships number in certain day, Q (d) _DownBe descending boats and ships number in certain day, Q (d) is the boats and ships number of up-downgoing in certain day;

The statistics monthly average volume of traffic or the annual volume of traffic just can obtain as long as the boats and ships number of up-downgoing every day in certain month, certain year added up to add up.Adopt excellent diagram form to represent among the present invention as a result after the statistics;

Described deadweight tonnage statistical module is used to add up hour deadweight tonnage, day deadweight capacity position, monthly average deadweight tonnage or annual deadweight tonnage by certain point on the navigation channel; Because the deadweight tonnage to the boats and ships that passed through among the present invention is estimated, therefore also can count hour deadweight tonnage by certain point on the navigation channel, day deadweight capacity position, monthly average day deadweight capacity position or annual deadweight tonnage very easily, hour deadweight tonnage can use formula (15) expression

$\stackrel{\_}{W}(i{)}_{\mathrm{up}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{W}_{\mathrm{up}}$

$\stackrel{\_}{W}(i{)}_{\mathrm{down}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{W}_{\mathrm{down}}$

$\stackrel{\_}{W}\left(i\right)=\stackrel{\_}{W}(i{)}_{\mathrm{up}}+\stackrel{\_}{W}(i{)}_{\mathrm{down}}$

In the formula: i is the integer between 0～23,, W (i) _UpBe up deadweight tonnage summation in certain i hour, W (i) _DownBe descending deadweight tonnage summation in certain i hour, W (i) is the deadweight tonnage summation of up-downgoing in certain i hour;

Day deadweight capacity position can use formula (16) to represent,

$\stackrel{\_}{W}(d{)}_{\mathrm{up}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}W(i{)}_{\mathrm{up}}$

$\stackrel{\_}{W}(d{)}_{\mathrm{down}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}W(i{)}_{\mathrm{down}}$

$\stackrel{\_}{W}\left(d\right)=\stackrel{\_}{W}(d{)}_{\mathrm{up}}+\stackrel{\_}{W}(d{)}_{\mathrm{down}}$

In the formula: W (d) _UpBe up boats and ships deadweight tonnage summation in certain day, W (d) _DownBe descending boats and ships deadweight tonnage summation in certain day,

Boats and ships deadweight tonnage summation for up-downgoing in certain day;

Statistics monthly average boats and ships deadweight tonnage summation or annual boats and ships deadweight tonnage summation just can obtain as long as the boats and ships deadweight tonnage summation of up-downgoing every day in certain month, certain year added up to add up.

Described traffic density statistical module is used to add up hour traffic density, daily traffic density, month traffic density or year traffic density by certain point on the navigation channel; Traffic density is meant a boats and ships number that travels in the unit length waters, it reflects the dense degree of boats and ships in this waters, so claim concentration of vessel again.Traffic density is an important parameters, the waters that traffic density is high, accident prone areas, the emphasis of vessel traffic research just and traffic administration naturally often.Traffic density can be used formula (17) expression,

$\mathrm{\&rho;}=\frac{N}{\mathrm{WL}}$ (17)

In the formula: ρ is the traffic density of certain moment; N is the boats and ships number; L is the length in observation navigation channel; W is the width in observation navigation channel.

Relational expression between the parameters such as the volume of traffic, traffic density, speed is called traffic flow model.The traffic flow basic model can be used formula (18) expression,

$\mathrm{\&rho;}=\frac{Q}{\mathrm{WV}}---\left(18\right)$

In the formula: ρ is a traffic density; Q is the volume of traffic; V is a speed; W is the width in observation navigation channel.

We are placed on W in the property file for width, the navigation channel coding (positional information of video observation station) in observation navigation channel, and the user can finish the input of these attribute datas on the attribute definition interface;

Because the headway to the boats and ships that passed through among the present invention estimates, hour ship's navigation average velocity can use formula (19) to represent,

$\stackrel{\_}{V}(i{)}_{\mathrm{up}}=\frac{1}{N}\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{V}_{\mathrm{up}}$ (19)

$\stackrel{\_}{V}(i{)}_{\mathrm{down}}=\frac{1}{N}\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{V}_{\mathrm{down}}$

$\stackrel{\_}{V}\left(i\right)=\frac{1}{2}(\stackrel{\_}{V}(i{)}_{\mathrm{up}}+\stackrel{\_}{V}(i{)}_{\mathrm{down}})$

In the formula: i is the integer between 0～23,, V (i) _UpBe the average speed per hour of up boats and ships in certain i hour, V (i) _DownBe the average speed per hour of descending boats and ships in certain i hour, V (i) is the average speed per hour of up-downgoing boats and ships in certain i hour; The traffic density that we can calculate hour according to formula (18), computing formula is provided by (20),

$\mathrm{\&rho;}(i{)}_{\mathrm{up}}=\frac{Q(i{)}_{\mathrm{up}}}{\stackrel{\_\_\_\_\_\_\_}{W*V}(i{)}_{\mathrm{up}}}---\left(20\right)$

$\mathrm{\&rho;}(i{)}_{\mathrm{down}}=\frac{\stackrel{\_}{Q}(i{)}_{\mathrm{down}}}{\stackrel{\_\_\_\_\_\_\_}{W*V}(i{)}_{\mathrm{down}}}$

$\mathrm{\&rho;}\left(i\right)=\frac{\stackrel{\_}{Q}\left(i\right)}{\stackrel{\_\_\_\_\_\_}{W*V}\left(i\right)}$

In the formula: i is the integer between 0～23,, ρ (i) _UpBe up navigation channel traffic density in certain i hour, ρ (i) _DownBe descending navigation channel traffic density in certain i hour, ρ (i) is certain i hour inner passage traffic density;

In a certain waters, channel span is fixed.After definite magnitude of traffic flow and these two parameters of headway, application of formula (20) can be extrapolated traffic density.But the above-mentioned traffic density of asking is an assembly average, is not to be the traffic density of certain moment.

Daily traffic density calculates can use formula (21) expression,

$\stackrel{\&OverBar;}{\mathrm{\&rho;}}{\left(d\right)}_{\mathrm{up}}=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}\mathrm{\&rho;}{\left(i\right)}_{\mathrm{up}}$ (21)

$\stackrel{\&OverBar;}{\mathrm{\&rho;}}{\left(d\right)}_{\mathrm{down}}=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}\mathrm{\&rho;}{\left(i\right)}_{\mathrm{down}}$

$\stackrel{\&OverBar;}{\mathrm{\&rho;}}\left(d\right)=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}\mathrm{\&rho;}\left(i\right)$

In the formula: ρ (d) _UpBe up navigation channel traffic density in certain day, ρ (d) _DownBe descending navigation channel traffic density in certain day, ρ (d) is a up-downgoing navigation channel traffic density in certain day;

Statistics monthly average navigation channel traffic density or annual navigation channel traffic density removed last corresponding fate then and just can be obtained as long as the navigation channel traffic density of up-downgoing every day in certain month, certain year added up to add up.

Embodiment 2

All the other are identical with embodiment 1, different is that large-range monitoring vision sensor 1 adopts the wide-angle imaging machine, as shown in Figure 1, wide-angle imaging machine and clipping the ball video camera are installed in a side same position in navigation channel, boats and ships in 140 ° of navigation channel scopes of wide-angle imaging function monitoring, the information fusion of same wide-angle imaging machine and clipping the ball video camera realizes by mapping table, each zone corresponding a preset point of fire ball video camera, measure boats and ships when certain surveyed area occurs in the wide-angle imaging machine examination, control fire ball video camera arrived and this area relative preset point after microprocessor obtained this information, and the feature of boats and ships object is captured.

Because the distance between the detection dummy line on the video image is represented with pixel, detection line has a funtcional relationship on this and the actual navigation channel, therefore need demarcate the wide-angle imaging machine, detect the corresponding relation between the detection line on dummy line and the navigation channel to set up; The demarcation of wide-angle imaging machine visual field distance relates to the theory of imaging geometry, the three-dimensional scenic on the actual navigation channel be projected the two-dimensional image plane of video camera, need set up the model of video camera and describe.Relation between its imaging system coordinate system has four coordinate systems as shown in Figure 16 among the figure, they are respectively: (1) actual coordinates, the three-dimensional scenic coordinate system XYZ on the just actual navigation channel; (2) camera coordinate system is the coordinate system xyz that formulate at the center with the wide-angle imaging machine; (3) photo coordinate system, formed photo coordinate system x ＇ y ＇ o ＇ in video camera; (4) computer picture coordinate system, the coordinate system MN that the computer-internal digital picture is used is a unit with the pixel.

The process that the wide-angle imaging machine is demarcated is exactly to derive the corresponding relation of computer two-dimensional image to three-dimensional scenic.Therefore the process of video camera imaging conversion is finished by three parts, the mutual conversion of respectively corresponding 4 coordinate systems.The precedence diagram of conversion as shown in Figure 15;

At first, (X, Y is Z) to camera coordinates (x, y, conversion z) from actual coordinate.Can adopt homogeneous coordinate system to represent if only consider the following conversion of situation of rigid body target, then transformation for mula can be expressed as by (22):

$\left[\begin{array}{c}\mathrm{<figure-callout\; id="1"\; label="x\; y\; z"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">x</figure-callout>}& \\ y\\ z\\ 1\end{array}\right]=R*\left[T*\left[\begin{array}{c}X\\ Y\\ Z\\ 1\end{array}\right]\right]---\left(22\right)$

Wherein, R is 4 * 4 rotation matrixs, and expression is as the rotational transform of plane in actual coordinates.Rotation can be represented by 3 Eulerian angle of Rigid Body in Rotation With, as shown in figure 17, can utilize Eulerian angle rotation matrix R can be expressed as the function of θ, φ and ψ

The Eulerian angle synoptic diagram as shown in Figure 17;

T is 4 * 4 translation matrix, represents to shift out as the plane initial point translational coordination of actual coordinate initial point, can be expressed as

$\left[\begin{array}{cccc}1& 0& 0& T_x\\ 0& 1& 0& T_y\\ 0& 0& 1& T_z\\ 0& 0& 0& 1\end{array}\right]$

With R, T substitution formula (22), obtain actual coordinate (X, Y, Z) to camera coordinates (x, y, transformation for mula z) (23),

$\left[\begin{array}{c}\mathrm{<figure-callout\; id="1"\; label="x\; y\; z"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">x</figure-callout>}& \\ y\\ z\\ 1\end{array}\right]=\left[\begin{array}{c}{r}_1(X+T_x)+r_2(Y+T_y)+r_3(Z+T_z)\\ r_4{(X+T_x)}_{}+r_5(Y+T_y)+r_6(Z+T_z)\\ r_7(X+T_x)+r_8(Y+T_y)+r_9(Z+T_z)\\ 1\end{array}\right]---\left(23\right)$

Then be from video camera three dimensional space coordinate (x, y, z) to the picture planimetric coordinates (x ＇, y ＇) conversion.Can get transformation for mula (24) according to the lens imaging principle,

$X\&prime;=f^{\frac{x}{z}}$ $y\&prime;=f^{\frac{y}{z}}$ (f<0) (24)

Be that picture planimetric coordinates (x ＇, y ＇) is to computer picture coordinate (M, conversion N) at last.The image coordinate unit that uses in the computing machine is the number of discrete pixel in the storer, so also need round the imaging plane that conversion just can be mapped to computing machine to reality as the coordinate on plane, its conversion expression formula is with formula (25) expression,

$M=O_m-\frac{x^{\&prime;}}{S_x}$ $N=O_n-\frac{y\&prime;}{S_y}---\left(25\right)$

Formula (23)～(25) mapping equation is combined, just can derive the computer picture coordinate (M, N) with actual three dimensional space coordinate (X, Y, transformation equation Z) (26),

$X=\frac{S_y(O_n-N)*f*({\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_3*r_8-{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">r</figure-callout>}}_2*r_9)+S_x(O_m-M)*f*({\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_5*r_9-{\mathrm{<figure-callout\; id="6"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_6*r_8)+f^2({\mathrm{<figure-callout\; id="6"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_6*r_2-{\mathrm{<figure-callout\; id="5"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_5*r_3)}{S_y(O_n-N)*f*({\mathrm{<figure-callout\; id="7"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_7*r_2-{\mathrm{<figure-callout\; id="8"\; label="r"\; filenames="A20071015638900391.png,A20071015638900411.png"\; state="\{\{state\}\}">r</figure-callout>}}_8*r_1)+S_x(O_m-M)*f*({\mathrm{<figure-callout\; id="8"\; label="r"\; filenames="A20071015638900391.png,A20071015638900411.png"\; state="\{\{state\}\}">r</figure-callout>}}_8*r_4-{\mathrm{<figure-callout\; id="7"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_7*r_5)+f^2({\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">r</figure-callout>}}_1*r_5-{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">r</figure-callout>}}_2*r_4)}(Z+T_z)-T_x$

$Y=\frac{S_y(O_n-N)*f*({\mathrm{<figure-callout\; id="9"\; label="r"\; filenames="A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_9*r_1-{\mathrm{<figure-callout\; id="7"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_7*r_3)+S_x(O_m-M)*f*({\mathrm{<figure-callout\; id="7"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_7*r_6-{\mathrm{<figure-callout\; id="9"\; label="r"\; filenames="A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_9*r_4)+f^2({\mathrm{<figure-callout\; id="3"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_3*r_4-{\mathrm{<figure-callout\; id="6"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_6*r_1)}{S_y(O_n-N)*f*({\mathrm{<figure-callout\; id="7"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_7*r_2-{\mathrm{<figure-callout\; id="8"\; label="r"\; filenames="A20071015638900391.png,A20071015638900411.png"\; state="\{\{state\}\}">r</figure-callout>}}_8*r_1)+S_x(O_m-M)*f*({\mathrm{<figure-callout\; id="8"\; label="r"\; filenames="A20071015638900391.png,A20071015638900411.png"\; state="\{\{state\}\}">r</figure-callout>}}_8*r_4-{\mathrm{<figure-callout\; id="7"\; label="r"\; filenames="A20071015638900391.png,A20071015638900401.png"\; state="\{\{state\}\}">r</figure-callout>}}_7*r_3)+f^2({\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="A20071015638900361.png,A20071015638900391.png"\; state="\{\{state\}\}">r</figure-callout>}}_1*r_5-{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="A20071015638900361.png,A20071015638900371.png"\; state="\{\{state\}\}">r</figure-callout>}}_2*r_4)}(Z+T_z)-T_y$

Z＝Z

In the formula: X, Y, Z are respectively the coordinate in the actual coordinates; Om, On represent the line number and the row at the some pixel place that the initial point as the plane is shone upon respectively on the computer picture plane; Sx, Sy represent the scale factor on computer picture plane x and the y direction respectively; F is the effective focal length of video camera; M, N are respectively the computer picture coordinates.

Embodiment 3

All the other are identical with embodiment 1, different is that large-range monitoring vision sensor 1 adopts the wide-angle imaging machine, as shown in Figure 5, the side that wide-angle imaging machine and clipping the ball video camera are installed in the navigation channel segment distance position that is separated by, the wide-angle imaging machine is taken direction and angle of direction along ng a path inclination, boats and ships in 140 ° of navigation channel scopes of wide-angle imaging function monitoring level stravismus, the information fusion of same wide-angle imaging machine and clipping the ball video camera realizes by mapping table, each zone corresponding a preset point of fire ball video camera, measure boats and ships when certain surveyed area occurs in the wide-angle imaging machine examination, control fire ball video camera arrived and this area relative preset point after microprocessor obtained this information, and the feature of boats and ships object is captured.

The above embodiments 1,2, the 3 invention effects that produced are: the master data that has realized automatic collection traffic on inland waters is also processed thereupon, analyze, various intelligent processing and statistical study, can be on the macroscopic view and microcosmic understand and grasp the actual state of cruiseway traffic, essential characteristic and universal law, traffic control for cruiseway, traffic administration, traffic programme, waterway network is built, supervision of law enforcement etc. provide a kind of intellectuality, the detection means of robotization, system also has ratio of performance to price height in addition, good reliability, safeguard simple, implement advantages such as convenient.

Claims (10)

1. the inner river ship automatic identification system that merges of a multiple vision sensor information, it is characterized in that: described inner river ship automatic identification system comprises the large-range monitoring vision sensor that monitors the navigation channel, carry out feature captures the fire ball vision sensor of hull image and name of vessel sign board image and is used for the microprocessor of boats and ships to image tracing, image recognition and traffic on inland waters situation statistics, described large-range monitoring vision sensor, fire ball vision sensor are connected with microprocessor, and described microprocessor comprises:

Image-display units is used for showing the video image of whole cruiseway, the general image of tracking boats and ships and the driving cabin image of following the tracks of boats and ships;

Large-range monitoring vision sensor calibration module is used to set up the navigation channel chart picture in space and the corresponding relation of the video image that is obtained;

The video data Fusion Module is used to control the rotation and the focusing of fire ball vision sensor, makes the fire ball vision sensor can aim at the tracking boats and ships and carries out the feature candid photograph;

The dummy line customized module, be used to customize the detection dummy line on the monitoring navigation channel, article three, detect that dummy line equates at interval and perpendicular to the navigation channel, article three, detect dummy line in the field range of large-range monitoring vision sensor, intersect a middle installation site of detecting dummy line and fire ball vision sensor;

Boats and ships enter and detect the dummy line detection module, are used for entering the detection dummy line if any boats and ships, produce an incident automatically, call corresponding processing module during incident of every generation automatically;

Boats and ships ID number and deposit the automatically-generating module of following the tracks of the boats and ships image folder, be used for the boats and ships of the dummy line of the visual field ragged edge that just enters the large-range monitoring vision sensor are named, and generate a file simultaneously with ID number name of these boats and ships, be used to deposit the close-up image of these boats and ships;

Module is captured in boats and ships general image location, is used to locate the image of the integral body of capturing these boats and ships;

The boats and ships profile detection module, be used to detect the edge contour of these boats and ships and estimate the height that this boats and ships hull surfaces, the notch cuttype edge detection algorithm of employing optimum-be that the canny edge detection algorithm carries out rim detection to the close-up image of this hull, obtain hull load-carrying feature and driving cabin characteristic information;

The multiple target tracking module is used to follow the tracks of the passing boats and ships on the navigation channel, the pixel of prospect boats and ships is extracted in the navigation channel background at the low-level image feature layer, and this tracking module comprises:

The adaptive background reduction unit, be used for cutting apart in real time dynamic object, the feature of using mixture gaussian modelling to come each pixel in the phenogram picture frame, when obtaining new picture frame, upgrade mixture gaussian modelling, on each time period, select the subclass of mixture gaussian modelling to characterize current background, if the pixel of present image and mixture gaussian modelling are complementary, judge that then this point is a background dot, otherwise judge that this point is the foreground point;

The connected region identify unit is used to extract prospect boats and ships object, adopts eight connected region extraction algorithms, utilize corrosion and expansion operator to remove isolated noise foreground point and the aperture of filling up the target area respectively, testing result is designated as { Rei, i=1,2,3 ... n}, the connected region that will detect gained at last projects on the initial prospect point set F again, gets communication with detection result { Ri=Rei ∩ F to the end, i=1,2,3,, n};

Tracking cell, be used for utilizing the color characteristic of ship target object to find the position and the size at motion ship target object place at video image, in the next frame video image, with moving target current position and big or small initialization search window, repeat this process and just can realize Continuous Tracking target;

Module is captured in location, ship-handling cabin, be used to locate the approximate location of the suspension name of vessel sign board of these boats and ships or lamp box and this position captured the image of the name of vessel sign trade mark, when tracked ship target object process detects dummy line, computing machine sends a series of instruction control fire ball vision sensors 2 automatically and turns to the residing orientation of ship target object, good focal length is captured then, the image of capturing has three at least, wherein an image is positive side position, and other two images are looked side ways the position about being respectively;

Name of vessel sign trade mark identification module is used to discern the name of vessel sign trade mark of shippping traffic, and this identification module comprises:

Based on the positioning unit of the name of vessel sign board of color model, be used for determining the position of name of vessel sign board at hull; Employing comes station keeping ship name sign board based on the LUV color model and the edge contour algorithm of even color space; Whether at first slightly judge has blueness or white to exist, if exist comparatively significantly blue portion or comparatively significantly white portion, jointing edge profile algorithm is determined the position of name of vessel sign board in the image of being captured;

The extraction unit of name of vessel sign board is used for the name of vessel sign board is extracted separately from hull, adopts the relation of complementary colors to extract the name of vessel sign board;

The correcting unit of name of vessel sign board, be used for that institute is extracted the rectification of name of vessel sign board image and be the front elevation picture, by the hough conversion rectangular coordinate is transformed into the parameter coordinate system, utilize K average fitting a straight line algorithm, detect two angles of ship board, utilize the method for " rotation+mistake is cut " conversion again, promptly earlier the angle of inclination is rotated and does to correct for the first time, carry out wrong shearization again and do rectification for the second time, deformation angle is carried out stretching conversion, whole licence plate is corrected into the licence plate of facing of a standard;

The Character segmentation unit of name of vessel sign board is used for the character on the name of vessel sign board after correcting is cut apart; Adopt the masterplate coupling each character on the name of vessel sign board to be cut apart in conjunction with the vertical and horizontal sciagraphy;

Character recognition unit, the character of each name of vessel sign board after being used for cutting apart is discerned, at first to carry out the character normalized, character is normalized to unified size, adopt the recognition methods of sorter for Chinese Character Recognition, the Chinese character after the normalized and standard Chinese character are mated discern; Adopt recognition methods for English alphabet, arabic numeral identification based on the BP neural network;

Boats and ships essential information retrieval module, being used for the name of vessel sign trade mark is major key, inquiry boats and ships essential information in the boats and ships Basic Information Table.

2. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 1 merges, it is characterized in that: described large-range monitoring vision sensor is the wide-angle vision sensor, is installed in a side of cruiseway.

3. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 1 merges, it is characterized in that: described large-range monitoring vision sensor is an omnibearing vision sensor, described omnibearing vision sensor is installed in the centre in navigation channel, described omnibearing vision sensor comprises the evagination catadioptric minute surface that is used for reflecting field, navigation channel object, evagination catadioptric minute surface down, be used to prevent anaclasis and the saturated dark circles cone of light, the dark circles cone is fixed on the center of catadioptric minute surface male part, be used to support the transparent cylinder of evagination catadioptric minute surface, be used to take the camera of imaging body on the evagination mirror surface, camera facing to the evagination mirror surface up.

4. as the inner river ship automatic identification system of the described multiple vision sensor information fusion of one of claim 1-3, it is characterized in that: described microprocessor also comprises:

Shipping draft and tonnage estimation block, be used to estimate the dead weight capacity of current boats and ships, the height that surfaces according to this boats and ships hull and leave long, high, wide data of these boats and ships in the boats and ships master database in, extrapolate this shipping draft, then the estimation dead weight capacity that calculates these boats and ships from the length and the width data of this ship; Estimation equation is: ship side is exposed at the height on the water surface during draft during Ship's Cargo draft=boats and ships ship side overall height-boats and ships hole capital after selling all securities-boats and ships loading; Water discharge (the metric ton)=wide * goods of long * draft * side's mode coefficient (cubic meter)/0.9756 (seawater) or 1 (fresh water) (cubic meter), displacement tonnage is as the deadweight tonnage of estimation boats and ships.

5. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 4 merges, it is characterized in that: described microprocessor also comprises:

The speed of the ship in metres per second detecting unit, be used to detect the headway of boats and ships, when boats and ships pass through to detect dummy line I respectively, when II and III, setting boats and ships ID is ShipID, initial tracing positional is StartPosition, initial time is StartTime, the end tracing positional is EndPosition, concluding time is EndTime), boats and ships ID enters at boats and ships and generates automatically when one of outermost detects dummy line, initial tracing positional is to enter position of detecting dummy line of outermost, initial time is to enter system time that detects dummy line of outermost, finishing tracing positional is to leave the monitoring field just to have run into position of detecting dummy line of outermost, and the concluding time is to leave the monitoring field just to have run into time of detecting dummy line of outermost, comes the headway of Ship ' by following formula:

$V\left(\mathrm{ShipID}\right)=\frac{\mathrm{EndPosition}-\mathrm{StartPosition}}{\mathrm{EndTime}-\mathrm{StartTime}}---\left(5\right).$

6. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 5 merges, it is characterized in that: described microprocessor also comprises:

Boats and ships up-downgoing judging unit is used to judge the navigation direction of boats and ships, if the boats and ships object of following the tracks of at first runs into is to detect dummy line I to be defined as up direction, is to detect dummy line III to be defined as down direction if the boats and ships object of following the tracks of at first runs into.

7. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 6 merges, it is characterized in that: described microprocessor also comprises:

Shippping traffic record generation module, be used for generating automatically a shippping traffic record, thereby traffic on inland waters amount statistics etc. performs data and prepares, and boat note comprises: boats and ships ID, the name of vessel sign trade mark, captain, the beam, height, estimation tonnage, detection time, speed of the ship in metres per second and up-downlink direction.

8. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 7 merges, it is characterized in that: described microprocessor also comprises:

Volume of traffic statistical module is used to add up hourly traffic volume, daily traffic volume, the monthly average volume of traffic or the annual volume of traffic by certain point on the navigation channel; The volume of traffic is meant the boats and ships number that passes through a certain waters in the unit interval, and the size of the volume of traffic directly reflects the busy extent of this waters vessel traffic; Because the volume of traffic is time dependent, so adopt hourly traffic volume, daily traffic volume, the monthly average volume of traffic or the annual volume of traffic to represent;

Hour the volume of traffic with formula (13) expression:

${\stackrel{\&OverBar;}{Q}\left(i\right)}_{\mathrm{up}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{Q}_{\mathrm{up}}---\left(13\right)$

${\stackrel{\&OverBar;}{Q}\left(i\right)}_{\mathrm{down}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{Q}_{\mathrm{down}}$

$\stackrel{\&OverBar;}{Q}\left(i\right)=\stackrel{\&OverBar;}{Q}{\left(i\right)}_{\mathrm{up}}+\stackrel{\&OverBar;}{Q}{\left(i\right)}_{\mathrm{down}}$

In the formula: i is the integer between 0～23,,

Be up boats and ships number in certain i hour, Be descending boats and ships number in certain i hour,

Boats and ships number for up-downgoing in certain i hour;

Daily traffic volume is represented with formula (14):

${\stackrel{\&OverBar;}{Q}\left(d\right)}_{\mathrm{up}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}{Q\left(i\right)}_{\mathrm{up}}---\left(14\right)$

${\stackrel{\&OverBar;}{Q}\left(d\right)}_{\mathrm{down}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}{Q\left(i\right)}_{\mathrm{down}}$

$\stackrel{\&OverBar;}{Q}\left(d\right)=\stackrel{\&OverBar;}{Q}{\left(d\right)}_{\mathrm{up}}+\stackrel{\&OverBar;}{Q}{\left(d\right)}_{\mathrm{down}}$

In the formula: Be up boats and ships number in certain day,

Be descending boats and ships number in certain day,

Boats and ships number for up-downgoing in certain day;

The statistics monthly average volume of traffic or the annual volume of traffic adds up as long as the boats and ships number of up-downgoing every day in certain month, certain year added up.

9. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 7 merges, it is characterized in that: described microprocessor also comprises:

The deadweight tonnage statistical module is used to add up hour deadweight tonnage, day deadweight capacity position, monthly average deadweight tonnage or annual deadweight tonnage by certain point on the navigation channel; Count hour deadweight tonnage by certain point on the navigation channel, day deadweight capacity position, monthly average day deadweight capacity position or annual deadweight tonnage;

Hour deadweight tonnage with formula (15) expression,

${\stackrel{\&OverBar;}{W}\left(i\right)}_{\mathrm{up}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{W}_{\mathrm{up}}$

${\stackrel{\&OverBar;}{W}\left(i\right)}_{\mathrm{down}}=\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{W}_{\mathrm{down}}---\left(15\right)$

$\stackrel{\&OverBar;}{W}\left(i\right)=\stackrel{\&OverBar;}{W}{\left(i\right)}_{\mathrm{up}}+\stackrel{\&OverBar;}{W}{\left(i\right)}_{\mathrm{down}}$

In the formula: i is the integer between 0～23,,

Be up deadweight tonnage summation in certain i hour,

Be descending deadweight tonnage summation in certain i hour,

Deadweight tonnage summation for up-downgoing in certain i hour;

Day deadweight capacity position can use formula (16) to represent,

${\stackrel{\&OverBar;}{W}\left(d\right)}_{\mathrm{up}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}{W\left(i\right)}_{\mathrm{up}}$

${\stackrel{\&OverBar;}{W}\left(d\right)}_{\mathrm{down}}=\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}{W\left(i\right)}_{\mathrm{down}}---\left(16\right)$

$\stackrel{\&OverBar;}{W}\left(d\right)=\stackrel{\&OverBar;}{W}{\left(d\right)}_{\mathrm{up}}+\stackrel{\&OverBar;}{W}{\left(d\right)}_{\mathrm{down}}$

In the formula:

Be up boats and ships deadweight tonnage summation in certain day,

Be descending boats and ships deadweight tonnage summation in certain day, Boats and ships deadweight tonnage summation for up-downgoing in certain day;

Statistics monthly average boats and ships deadweight tonnage summation or annual boats and ships deadweight tonnage summation adds up as long as the boats and ships deadweight tonnage summation of up-downgoing every day in certain month, certain year added up.

10. the inner river ship automatic identification system that multiple vision sensor information as claimed in claim 7 merges, it is characterized in that: described microprocessor also comprises:

The traffic density statistical module is used to add up hour traffic density, daily traffic density, month traffic density or year traffic density by certain point on the navigation channel; Traffic density is meant a boats and ships number that travels in the unit length waters, it reflects the dense degree of boats and ships in this waters, traffic density formula (17) expression,

$\mathrm{\&rho;}=\frac{N}{\mathrm{WL}}---\left(17\right)$

In the formula: ρ is the traffic density of certain moment; N is the boats and ships number; L is the length in observation navigation channel; W is the width in observation navigation channel.

Relational expression between the parameters such as the volume of traffic, traffic density, speed is called traffic flow model, formula (18) expression of traffic flow basic model,

$\mathrm{\&rho;}=\frac{Q}{\mathrm{WV}}---\left(18\right)$

In the formula: ρ is a traffic density; Q is the volume of traffic; V is a speed; W is the width in observation navigation channel.

Ship's navigation average velocity hourly is represented with formula (19):

${\stackrel{\&OverBar;}{V}\left(i\right)}_{\mathrm{up}}=\frac{1}{N}\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{V}_{\mathrm{up}}---\left(19\right)$

${\stackrel{\&OverBar;}{V}\left(i\right)}_{\mathrm{down}}=\frac{1}{N}\underset{n=i}{\overset{n=i+1}{\mathrm{\&Sigma;}}}{V}_{\mathrm{down}}$

$\stackrel{\&OverBar;}{V}\left(i\right)=\frac{1}{2}(\stackrel{\&OverBar;}{V}{\left(i\right)}_{\mathrm{up}}+\stackrel{\&OverBar;}{V}{\left(i\right)}_{\mathrm{down}})$

In the formula: i is the integer between 0～23,, Be the average speed per hour of up boats and ships in certain i hour,

Be the average speed per hour of descending boats and ships in certain i hour, V (i) is the average speed per hour of up-downgoing boats and ships in certain i hour; The traffic density that we can calculate hour according to formula (18), computing formula is provided by (20),

${\mathrm{\&rho;}\left(i\right)}_{\mathrm{up}}=\frac{\stackrel{\&OverBar;}{Q}{\left(i\right)}_{\mathrm{up}}}{{\stackrel{\&OverBar;}{W*V}\left(i\right)}_{\mathrm{up}}}---\left(20\right)$

$\mathrm{\&rho;}{\left(i\right)}_{\mathrm{down}}=\frac{\stackrel{\&OverBar;}{Q}{\left(i\right)}_{\mathrm{down}}}{\stackrel{\&OverBar;}{W*V}{\left(i\right)}_{\mathrm{down}}}$

$\mathrm{\&rho;}\left(i\right)=\frac{\stackrel{\&OverBar;}{Q}\left(i\right)}{\stackrel{\&OverBar;}{W*V}\left(i\right)}$

In the formula: i is the integer between 0～23,, ρ (i) _UpBe up navigation channel traffic density in certain i hour, ρ (i) _DownBe descending navigation channel traffic density in certain i hour, ρ (i) is certain i hour inner passage traffic density;

In a certain waters, channel span is fixing, and after definite magnitude of traffic flow and these two parameters of headway, application of formula (20) is extrapolated traffic density;

Daily traffic density calculates with formula (21) expression,

${\stackrel{\&OverBar;}{\mathrm{\&rho;}}\left(d\right)}_{\mathrm{up}}=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}\left(i\right)}_{\mathrm{up}}$

${\stackrel{\&OverBar;}{\mathrm{\&rho;}}\left(d\right)}_{\mathrm{down}}=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}{\mathrm{\&rho;}\left(i\right)}_{\mathrm{down}}---\left(21\right)$

$\stackrel{\&OverBar;}{\mathrm{\&rho;}}\left(d\right)=\frac{1}{24}\underset{i=0}{\overset{i=23}{\mathrm{\&Sigma;}}}\mathrm{\&rho;}\left(i\right)$

In the formula:

Be up navigation channel traffic density in certain day, Be descending navigation channel traffic density in certain day,

Be up-downgoing navigation channel traffic density in certain day;

Statistics monthly average navigation channel traffic density or annual navigation channel traffic density removed last corresponding fate then as long as the navigation channel traffic density of up-downgoing every day in certain month, certain year added up to add up.

Images