CN101034005A - Switching method between videos for detecting vehicles during day-night - Google Patents
Switching method between videos for detecting vehicles during day-night Download PDFInfo
- Publication number
- CN101034005A CN101034005A CN 200710039378 CN200710039378A CN101034005A CN 101034005 A CN101034005 A CN 101034005A CN 200710039378 CN200710039378 CN 200710039378 CN 200710039378 A CN200710039378 A CN 200710039378A CN 101034005 A CN101034005 A CN 101034005A
- Authority
- CN
- China
- Prior art keywords
- brightness
- intermediate value
- sunrise
- sunset
- night
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
This invention relates to a kind of day and night video vehicle detecting compartment's switch method, belong to aptitude trafficking technosphere. The invention by setting operational area that is able to cover whole road's each driveway, to obtain attributive information of road's variations brightness velocity form digital image, and by this to judge when processing day and night test method compartment's switch. In addition , this method divide one day into early morning early morning, day, evening and nighttime four stage; take sunrise and sunset time as reference point to set sunrise and sunset two transient time section; the switch method only function at transient time section, by multi- sensor's complementarily information confluent strategy, to solve problem of two vehicle instrumentation method's low detecting rate at day and night transient time section.
Description
Technical field
What the present invention relates to is a kind of method of technical field of traffic transportation vehicle detection, specifically is the changing method between a kind of video frequency vehicle round the clock detects.
Background technology
Along with expanding economy, the raising of living standards of the people, the speed that the growth rate of automobile quantity is built considerably beyond road infrastructure, bus or train route contradiction becomes increasingly conspicuous.Intelligent transportation system (ITS) has been acknowledged as and has solved problems such as road traffic congestion, accident be frequent, ensures the smooth and easy and safe effective way of traffic.Vehicle detection is the basic key problem of intelligent transportation system.Compare with traditional detecting device, but the video frequency vehicle detecting device have the convenient and lower monitoring range of expense of installation and maintenance wide, can obtain plurality of advantages such as a greater variety of traffic parameters, thereby in intelligent transportation system, obtained using widely.But the testing result of video sensor is subjected to the influence of factors such as climatic environment and intensity of illumination bigger.Particularly daytime and night, video image processing system must adapt to this environmental change round the clock.Therefore the video frequency vehicle detection system of round-the-clock running is between daytime and these two great periods of road lighting difference at night, usually adopt two kinds of different vehicle checking methods to guarantee that it detects accuracy rate respectively, and the suitable time is switched this two kinds of methods to adopt certain policy selection.
Through the literature search of prior art is found, present existing automobile video frequency detection system mainly comprises the fixedly onboard system of traffic information acquisition system and Seiichi NAGUMO such as Peek, Iteris, Autoscope, Setrix.They select the method for switching time to have two kinds usually: a kind of method is to obtain ambient light according to information from imaging devices such as video camera, determine switching time as Setrix with the CMOS automatic controlled variable of exposing, but this method is had relatively high expectations for imaging device, causes the cost of system to rise.Another kind method is to obtain illumination information from digital picture.Seiichi Nagumo etc. are at " electronics and computer engineering " IEEE CCECE2003 meeting collection, " the Extraction offorward vehicles by front-mounted camera using brightness information " that delivers on rolling up the 1243rd to 1246 page May the 2nd in 2003, (extraction of the preposition camera driver vehicle by using illumination information), propose in this article to obtain illumination information from digital picture, concrete grammar is: utilize onboard system to differentiate round the clock according to the chromatic information of image the first half sky; The digital picture of gathering is divided into two, threshold value of the first half setting is carried out color decision decide round the clock.Its deficiency is: for the fixed video detection system, its camera generally is installed on road bar or other roadside buildings following road conditions is monitored, can't obtain the image information of sky; And,, no matter take any of two kinds of detection methods because intensity of illumination sharply changes in transition period round the clock (early morning and at dusk), its accuracy rate all can reduce (being that method lost efficacy) greatly, has influence on the decision-making and the scheduling of upper strata traffic control system.Must employing address this problem someway, rather than at a time carry out the method switching simply.This is the problem that solved of the present invention just also.
Summary of the invention
The objective of the invention is at the deficiencies in the prior art, needs according to the fixed video vehicle detecting system, propose a kind of based on the changing method between the detection of video frequency vehicle round the clock of road surface brightness intermediate value rate of change, make it can directly from digital picture, obtain road brightness change information, and carry out the switching between detection method round the clock on this basis, imaging device there is not special requirement.The present invention has that computation complexity is low, real-time good, switches effect advantage accurately.
The present invention is achieved by the following technical solutions: be divided in one day daytime, at dusk, night and early morning four working stages.Be the time that vehicle checking method is effectively worked between daytime daytime, and be the method operational use time at night night.Dusk and early morning then are between daytime and the time period that night, method all lost efficacy, video system detects information fusion by carrying out complementarity with other magnetosensitive detecting device that not influenced by illumination variation in this two periods, eliminates the low excessively influence of verification and measurement ratio that the detection method of video system own lost efficacy and brings.Therefore for the handoff procedure from the daytime to the night, carried out in fact from daytime at dusk with from dusk to twice switching at night; Similarly, for the process from the night to the daytime, also carried out twice switching from the night to the early morning and from the early morning to the daytime; For changing method round the clock, needn't whole day all move in 24 hours.The present invention is a benchmark with sunrise time point and sunset time point, being divided into night, sunrise period, daytime, sunset period and night in one day 24 hours, set up sunrise (sunrise, from the night to the daytime) and sunset (sunset, from the daytime to the night) two transit time section, the sunrise period comprises night, early morning and daytime, and the sunset period comprises daytime, dusk and night.Sunrise and at sunset segment length be 1 hour, thereby can cover at dusk stage and stage in early morning, every day is only at these two transition period execution switching methods.The changing method that the present invention proposes, the main foundation whether decision switches is exactly the intensity of illumination variation characteristic on road surface.When a new brightness intermediate value sample is read in system, changed 10 if find the brightness intermediate value, with regard to a newly-built broken line.Because the transformation period length T of every line segment, i.e. this line segment brightness changed for 10 used times, was inversely proportional to brightness intermediate value pace of change in this section period.When the switching judging rule switches if just deciding according to this pace of change characteristic information of time span T.Descending or ascending rapid variation has taken place in the T value of finding newly-generated line segment in the transition period, just determines whether carrying out the switching of detection method round the clock according to the switching judging rule.
During practical operation of the present invention, below video image, get a zone that can cover each track of entire road but not exceed the road scope as far as possible, be called the operation area, weigh road surface intensity of illumination and situation of change thereof by the intermediate value of calculating all pixel brightness values in the operation area, with as the decision-making feature foundation of whether switching; Present ascendant trend in road section surface brightness intermediate value at sunrise in time from plateau, the state that tends to be steady again at last presents downtrending in road section surface brightness intermediate value at sunset in time from plateau, and state at last tends to be steady again; They have common characteristic: promptly trend all roughly is the serpentine shape, all compares slowly in the brightness variation of beginning and ending phase.And in the stage casing of " S ", the zone between two flex points then exists the process that a brightness changes (descend or rise) fast, and brightness sharply changes in this process, between daytime and night the detection method accuracy rate all become very low, corresponding to early morning and stage at dusk.The present invention seeks two flex points as prerequisite according to the pace of change information of method operation area brightness intermediate value, and finishes twice switching of transition period at the flex point near zone.
The present invention includes following step:
The first step, determine road surface brightness and signal Synchronization:
When sunrise and sunset begin, take wagon flow on the highway by framework at the video camera on the support, the brightness on road surface can reflect the variation of weather and replacement round the clock in the video.Therefore accurately obtaining monochrome information is a crucial step, and the brightness intermediate value can be described the variation of illumination preferably, if but computing brightness intermediate value can not satisfy real-time requirement because calculated amount is big.In order to reduce calculated amount, guarantee the real-time of changing method, the present invention is not the brightness intermediate value that every two field picture all calculates the operation area, but per second calculates once.For reaching tens of minutes switching transient process, 1 time/second sampling rate can guarantee that gained intermediate value time series accurately reflects the situation of change of real road brightness.Simultaneously, rack-mount magneto-dependent sensor also begins to gather waveform signal, carries out vehicle detection, performs with video and synchronouss working.
Second step, brightness pre-service and signal denoising:
After video image acquisition brightness intermediate value, also to carry out pre-service.Because the influence that crossed by environmental change and vehicle, there is more burr in the original intermediate value sequence curve that the first step obtained, to the bigger influence of extraction existence of brightness pace of change characteristic information, therefore will carry out smothing filtering to eliminate burr to it.The present invention adopts the approximation method of mean filter: the accumulative total method of average, i.e. the value of this filtering equals the weighted sum of the intermediate value of last filter value and this new acquisition.Obtain level and smooth curve after the pre-service of brightness intermediate value and be used for subsequent analysis processing.For the magnetosensitive signal, also to carry out signal Processing, eliminate noise.
In the 3rd step, the brightness variation characteristic extracts and signal threshold value is chosen:
For the intermediate value curve after level and smooth, need to extract brightness pace of change characteristic information, switching for method provides basis for estimation.The present invention adopts PLR (piecewise linearity is represented) method (Piecewise LinearRepresentation).When a new brightness intermediate value sample is read in system, changed 10 if find the brightness intermediate value, with regard to a newly-built broken line.The starting point of this line segment is the terminal point of last broken line, and its terminal point then is current brightness intermediate value.Thereby with the linearization of whole piece brightness median smoothing curve segmentation.The transformation period length T of every line segment, i.e. this line segment brightness changed for 10 used times, was inversely proportional to brightness intermediate value pace of change in this section period, can be used as brightness intermediate value pace of change characteristic information.By piece-wise linearization, on the one hand can extract brightness pace of change information effectively, because the 4th step switching judging is only carried out when generating new line segment basically, so PLR can reduce the judgement number of times of switching, reduces the expense of system on the other hand.The magnetosensitive signal is carried out selection of threshold so that determine whether to be signals of vehicles equally.
In the 4th step, switching principle switches and information fusion in accordance with regulations:
For reducing calculated amount, satisfy the real-time requirement that method is switched, the present invention basically only each generate new broken line in, determine whether switching according to this pace of change characteristic information of time span T of this line segment.When just entering transition period, the brightness intermediate value changes slowly, and the time span T of the line segment that generates is roughly about 15 minutes; When the T value of finding newly-generated line segment is about 1 minute, show the quickening of brightness intermediate value pace of change, entered first flex point region, just carry out switching the first time; When entering second flex point zone, pace of change slows down afterwards, and it is big that the T value of newly-generated line segment becomes, and then carries out switching the second time, thereby finishes the whole handoff procedure of method round the clock.Two transition periods of sunrise and sunset are all followed such switching principle, and the sunrise period switches according to the order in " night → early morning → daytime ", and the sunset period then switches according to the order on " daytime → at dusk → night ".1 hour transition period in sunrise and sunset all adopts video and magneto-dependent sensor information fusion method to carry out vehicle detection.The process of information fusion is as follows::
1. calculate credible substantially number, belief function and the likelihood degree function of two sensors respectively;
2. utilize Dempster to merge rule, try to achieve credible substantially number, belief function and likelihood degree function under all the sensors synergy;
3. under the decision rule of setting, select to have the target of maximum support.
The present invention with time of one day be divided into early morning, daytime, dusk and night four-stage, by adopting the strategy that carries out information fusion in the morning with and other magneto-dependent sensor that not influenced by illumination variation, solved the lost efficacy problem of the verification and measurement ratio reduction that brings of two kinds of vehicle checking methods of transition period round the clock in the stage the dusk.Magneto-dependent sensor carries out vehicle detection and mainly utilizes vehicle to cause the changes of magnetic field around the sensor and detect vehicle through out-of-date meeting.The ultimate principle of multi-sensor information fusion technology also makes full use of a plurality of sensor resources just as human brain overall treatment information, by reasonable domination and use to these sensors and observation information thereof, redundancy or the complementary information of a plurality of sensors on time and space made up according to certain criterion, explain or description with the consistance of obtaining object being observed.The elementary object of data fusion is to derive more effective informations by data-optimized combination.Its final purpose is to utilize the advantage of the common or joint operation of a plurality of sensors, improves the validity of a plurality of sensing systems.
Compared with prior art, it is low that the present invention has a computation complexity, selects advantage accurately switching time, is fit to the real-time video vehicle detection and uses.And by daytime, at dusk, night and early morning four-stage division and carry out the strategy of the complementary information fusion of multisensor between the lights with the stage in early morning, solved video sensor and be subjected to the influence of illumination variation and the lower problem of verification and measurement ratio in transition period.The complementary information of multisensor is sensor informations such as comprehensive video, magnetosensitive.Magneto-dependent sensor utilizes the magnetic field induction vehicle, not influenced by other environmental factors such as illumination, wind and rain etc.; Between the lights with 1 hour transition period in early morning, because light changes greatly, utilize the characteristics of magneto-dependent sensor, based on magneto-dependent sensor, be aided with the additional mutually fusion that video carries out information, by calculating both probability density, give different weights and embody shared separately importance and carry out information fusion, further improve the vehicle detection accuracy rate.
Description of drawings
The synoptic diagram that Fig. 1 is provided with for transition period of the present invention, wherein: Fig. 1 (a) is the transition period setting; Fig. 1 (b) is the signal of sunrise period; Fig. 1 (c) is the signal of sunset period.
The synoptic diagram that Fig. 2 chooses for the changing method operation area, wherein Fig. 2 (a) is the sunrise period; Fig. 2 (b) is the sunset period.
Fig. 3 is the process flow diagram based on the changing method of road surface brightness intermediate value rate of change.
Fig. 4 is the original brightness intermediate value time-serial position that sampling rate obtained of whole transition period by 1 frame/second, and wherein Fig. 4 (a) is the sunrise period; Fig. 4 (b) is the sunset period.
Fig. 5 is the curve of original brightness intermediate value curve behind accumulative total method of average smothing filtering, and wherein Fig. 5 (a) is the sunrise period; Fig. 5 (b) is the sunset period.
Fig. 6 carries out the synoptic diagram that the PLR piecewise linearity is represented for brightness median smoothing curve, and wherein Fig. 6 (a) is the sunrise period; Fig. 6 (b) is the sunset period.
Fig. 7 illustrates that to the process that the sample video carries out detection method switching round the clock wherein Fig. 7 (a) is the sunrise period for the present invention; Fig. 7 (b) is the sunset period.The sunrise period switches according to the order in " night → early morning → daytime ", and the sunset period then switches according to the order on " daytime → at dusk → night ".
Embodiment
Below in conjunction with accompanying drawing embodiments of the invention are elaborated: present embodiment is being to implement under the prerequisite with the technical solution of the present invention, provided detailed embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment.
The sample video acquisition that this example adopted is in A4 expressway, Shanghai, comprises sunrise period sample from the night to the daytime and the sunset period sample from the daytime to the night.With sunrise time point and sunset time point is benchmark, being divided into night, sunrise period, daytime, sunset period and night in one day 24 hours, set up sunrise (sunrise, from the night to the daytime) and sunset (sunset, from the daytime to the night) two transit time section, sunrise and at sunset segment length be 1 hour, thereby can cover stage and stage in early morning at dusk, every day, the order that sunrise and sunset are switched was respectively shown in Fig. 1 (b) and Fig. 1 (b) only at these two transition period execution switching methods such as Fig. 1 (a).Choosing respectively shown in Fig. 2 (a) and Fig. 2 (b) of sunrise and sunset changing method operation area.During execution switching method, road surface intensity of illumination and situation of change thereof are weighed by the intermediate value of calculating all pixel brightness values in the operation area by system, with as the decision-making feature foundation of whether switching; The setting of sunrise and sunset transition period is a benchmark with sunrise and sunset time point respectively then.The sunrise period started from the sunrise time point preceding 40 minutes, ended at sunrise and put back 20 minutes.The sunset period started from sunset preceding 20 minutes, ended at post sunset 40 minutes.The sun set/raise time point of every day can be found from relevant uranology data.
In the transition period handoff procedure, when system was in early morning or in the stage the dusk, the verification and measurement ratio with the vehicle at night detection method between daytime was all very low.Magneto-dependent sensor has the advantage that testing result is not subjected to factor affecting such as weather, illumination, therefore the present invention adopts between video detector and magnetosensitive detecting device based on the complementary scheme that merges of the multi-sensor information of DS argument, carry out multi-sensor information fusion by the probability density function that calculates two class sensors, thereby remedied the deficiency of video detector, guaranteed transition period vehicle detection result's accuracy.
As shown in Figure 3, changing method of the present invention comprises following step:
1, determine road surface brightness and signal Synchronization:
When sunrise and sunset begin, take wagon flow on the highway by framework at the video camera on the support, the brightness on road surface can reflect the variation of weather and replacement round the clock in the video.Therefore accurately obtaining monochrome information is a crucial step, and the brightness intermediate value can be described the variation of illumination preferably, if but computing brightness intermediate value can not satisfy real-time requirement because calculated amount is big.In order to reduce calculated amount, guarantee the real-time of changing method, this method is not the brightness intermediate value that every two field picture all calculates the operation area, but per second calculates once.For reaching tens of minutes switching transient process, 1 time/second sampling rate can guarantee that gained intermediate value time series accurately reflects the situation of change of real road brightness.Fig. 4 (a) and Fig. 4 (b) are respectively the sample brightness intermediate value time-serial position signal of sunrise and two transition periods of sunset, and curve shape meets " S " type substantially.Simultaneously, rack-mount magneto-dependent sensor also begins to gather waveform signal, carries out vehicle detection, performs with video and synchronouss working.
2, brightness pre-service and signal denoising:
After video image acquisition brightness intermediate value, also to carry out pre-service.Because the influence that crossed by environmental change and vehicle, there is more burr in the original intermediate value sequence curve that the first step obtained, to the bigger influence of extraction existence of brightness pace of change characteristic information, therefore it is carried out smothing filtering to eliminate burr.The present invention adopts the accumulative total method of average, and the value of this filtering equals the weighted sum of the intermediate value of last filter value and this new acquisition, i.e. avr
j=(1-α) avr
J-1+ α x
j, avr is the mean filter value, x
jIt is the new samples value.α is a update coefficients.It is low that this method has a computation complexity, the noise remove good effectiveness.Shown in Fig. 5 (a) and Fig. 5 (b), burr is substantially by filtering respectively for level and smooth result of sunrise and sunset period.Obtain level and smooth curve after the pre-service of brightness intermediate value and be used for subsequent analysis processing.For the magnetosensitive signal, also to carry out signal Processing, eliminate noise.
3, extraction of brightness variation characteristic and signal threshold value are chosen:
For the intermediate value curve after level and smooth, need to extract brightness pace of change characteristic information, switching for method provides basis for estimation.The present invention adopts PLR (piecewise linearity is represented) method.When a new brightness intermediate value sample is read in system, changed 10 if find the brightness intermediate value, with regard to a newly-built broken line.The starting point of this line segment is the terminal point of last broken line, and its terminal point then is current brightness intermediate value.Thereby with the linearization of whole piece brightness median smoothing curve segmentation.Sunrise and sunset period PLR represent that respectively the longitudinal axis represent brightness intermediate value shown in Fig. 6 (a) and figure (b), and the brightness that the projection of line segment on the longitudinal axis is this line segment changes big or small, and unification is 10; Transverse axis is represented the time, and unit is second, and the projection of line segment on transverse axis is this line segment brightness and changed for 10 used times, is called the transformation period length T of line segment, is inversely proportional to brightness intermediate value pace of change in this section period.The time span of different line segments is different, can be used as brightness intermediate value pace of change characteristic information during this period of time.The magnetosensitive signal is carried out selection of threshold so that determine whether to be signals of vehicles equally.
4, switching principle switches and information fusion in accordance with regulations:
For reducing calculated amount, satisfy the real-time requirement that method is switched, except that final stage, the present invention only each generate new broken line in, determine whether switching according to this pace of change characteristic information of time span T of this line segment.1 hour transition period in sunrise and sunset all adopts video and magneto-dependent sensor information fusion method to carry out vehicle detection.
The sunrise period switches according to the order in " night → early morning → daytime ", and its judgment rule is as follows:
A) in the stage at night, brightness intermediate value originally changes slowly.When the T value of finding newly-generated line segment sharply diminishes, show the pace of change quickening, entered first flex point region, just carry out switching the first time.If but intermediate value has been higher than luminance threshold 40 before this, then road surface brightness becomes big, and method accuracy at night descends, and just switches by force.
B) in the morning in the stage, the brightness intermediate value is in the fast rise stage, preserves the T value of each newly-generated line segment.If intermediate value has been higher than luminance threshold 90, the sky brightens basically fully, and the condition of method operation begins to possess between daytime.Calculate the average value mu and the standard deviation sigma of all line segment T values of having preserved this moment.If after this do not generate new line segment yet if surpass the time of max (2 μ, μ+3 σ), illustrate the intermediate value pace of change faster ascent stage slow down, enter second flex point zone, just carry out the switching second time; If the switching in night to early morning is forced to carry out, then wait to enter quick changes phase and carry out above-mentioned rule again.
C) by day in the stage, system's vehicle checking method between daytime that brings into operation need not to do any switching judging.But,, guarantee the carrying out of system's vehicle detection between daytime then force to carry out to switch the second time just in case, remain unfulfilled the switching on early morning to daytime until sunrise transition period ends with system because some is special.
The sunset period then switches according to the order on " daytime → at dusk → night ".Concrete judgment rule is the same substantially with the sunrise stage, and just brightness intermediate value change procedure becomes decline by rising, and the luminance threshold that switches for the first time and switch for the second time becomes 100 and 50 respectively.The synoptic diagram that Fig. 7 (a) and Fig. 7 (b) are sunrise respectively and sunset transition period utilization the present invention carries out method secondary switching round the clock.As seen the present invention can automatically finish the accurate switching of vehicle checking method round the clock basically.
Claims (4)
1, changing method between a kind of video frequency vehicle round the clock detects, it is characterized in that: with sunrise time point and sunset time point is benchmark, being divided into night in one day 24 hours, the sunrise period, daytime, sunset period and night, set up sunrise and sunset section two transit times, the sunrise period comprises night, early morning and daytime, the sunset period comprises daytime, dusk and night, sunrise and at sunset segment length be 1 hour, thereby can cover stage and stage in early morning at dusk, every day, the foundation of switching was exactly the intensity of illumination variation characteristic on road surface only at these two transition period execution switching methods; During practical operation, below video image, get a zone that can cover each track of entire road but not exceed the road scope, be called the operation area, weigh road surface intensity of illumination and situation of change thereof by the intermediate value of calculating all pixel brightness values in the operation area, with as the decision-making feature foundation of whether switching; Present ascendant trend in road section surface brightness intermediate value at sunrise in time from plateau, state at last tends to be steady again, present downtrending in road section surface brightness intermediate value at sunset in time from plateau, state at last tends to be steady again, they have common characteristic: promptly trend all roughly is the serpentine shape, brightness variation in beginning and ending phase is all compared slowly, and in the stage casing of " S ", zone between two flex points, then exist the process that a brightness descends or rises fast, corresponding to early morning and in the stage the dusk, as prerequisite, seek two flex points according to the pace of change information of operation area brightness intermediate value, and finish twice switching of transition period at the flex point near zone.
2, the changing method between video frequency vehicle round the clock according to claim 1 detects is characterized in that, comprises the steps:
The first step, determine road surface brightness and signal Synchronization:
When sunrise and sunset begin, by the wagon flow of framework on the shooting of the video camera on support highway, the brightness on road surface can reflect the variation of weather and replacement round the clock in the video, per second calculates the brightness intermediate value of an operation area image, simultaneously, adopt rack-mount magneto-dependent sensor to gather waveform signal, carry out vehicle detection, perform with video and synchronous working;
Second step, brightness pre-service and signal denoising:
After video image acquisition brightness intermediate value, also to carry out pre-service, promptly carry out smothing filtering to eliminate burr, obtain level and smooth curve after the pre-service of brightness intermediate value and be used for subsequent analysis processing, also carry out signal Processing for the magnetosensitive signal, eliminate noise;
In the 3rd step, the brightness variation characteristic extracts and signal threshold value is chosen:
For the intermediate value curve after level and smooth, extract brightness pace of change characteristic information, for the method switching provides basis for estimation, adopt the piecewise linearity method for expressing: when a new brightness intermediate value sample is read in system, changed 10 if find the brightness intermediate value, with regard to a newly-built broken line, the starting point of this line segment is the terminal point of last broken line, its terminal point then is current brightness intermediate value, thereby with the linearization of whole piece brightness median smoothing curve segmentation; The transformation period length T of every line segment, i.e. this line segment brightness changed for 10 used times, be inversely proportional to brightness intermediate value pace of change in this section period,, the magnetosensitive signal carried out selection of threshold so that determine whether to be signals of vehicles equally as brightness intermediate value pace of change characteristic information;
In the 4th step, switching principle switches and information fusion in accordance with regulations:
Only each generate new broken line in, determine whether switching according to this pace of change characteristic information of time span T of this line segment, when just entering transition period, the brightness intermediate value changes slowly, the time span T of the line segment that generates is 15 minutes; When the T value of finding newly-generated line segment during at 1 minute, show that brightness intermediate value pace of change accelerates, entered first flex point region, just carry out the switching first time; When entering second flex point zone, pace of change slows down afterwards, and it is big that the T value of newly-generated line segment becomes, and then carries out switching the second time, thereby finishes the whole handoff procedure of method round the clock; Two transition periods of sunrise and sunset are all followed such switching principle, the sunrise period switches according to the order in " night → early morning → daytime ", the sunset period then switches according to the order on " daytime → at dusk → night ", all adopts video and magneto-dependent sensor information fusion method to carry out vehicle detection in 1 hour transition period of sunrise and sunset.
3, the changing method between video frequency vehicle round the clock according to claim 2 detects, it is characterized in that, the described smothing filtering that carries out, the approximation method of employing mean filter: the accumulative total method of average, i.e. the value of this filtering equals the weighted sum of the intermediate value of last filter value and this new acquisition.
4, the changing method between video frequency vehicle round the clock according to claim 2 detects is characterized in that, described information fusion, and process is as follows:
1. calculate credible substantially number, belief function and the likelihood degree function of video and magneto-dependent sensor respectively;
2. utilize Dempster to merge rule, try to achieve credible substantially number, belief function and likelihood degree function under all the sensors synergy;
3. under the decision rule of setting, select to have the target of maximum support.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100393782A CN100504318C (en) | 2007-04-12 | 2007-04-12 | Switching method between videos for detecting vehicles during day-night |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100393782A CN100504318C (en) | 2007-04-12 | 2007-04-12 | Switching method between videos for detecting vehicles during day-night |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101034005A true CN101034005A (en) | 2007-09-12 |
| CN100504318C CN100504318C (en) | 2009-06-24 |
Family
ID=38730661
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2007100393782A Expired - Fee Related CN100504318C (en) | 2007-04-12 | 2007-04-12 | Switching method between videos for detecting vehicles during day-night |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100504318C (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101727748A (en) * | 2009-11-30 | 2010-06-09 | 北京中星微电子有限公司 | Method, system and equipment for monitoring vehicles based on vehicle taillight detection |
| CN102044152A (en) * | 2010-11-19 | 2011-05-04 | 杭州海康威视系统技术有限公司 | Day and night video detecting method and device |
| CN102779272A (en) * | 2012-06-29 | 2012-11-14 | 惠州市德赛西威汽车电子有限公司 | Switching method for vehicle detection modes |
| CN102842025A (en) * | 2011-06-25 | 2012-12-26 | 北京汉王智通科技有限公司 | Detection scene judgment method and detection scene judgment device of video image |
| CN102881162A (en) * | 2012-09-29 | 2013-01-16 | 北京市交通信息中心 | Data processing and fusion method for large-scale traffic information |
| CN105187731A (en) * | 2015-09-18 | 2015-12-23 | 中电海康集团有限公司 | Day-night switching method and device based on forward collision warning |
| CN105389991A (en) * | 2015-12-03 | 2016-03-09 | 杭州中威电子股份有限公司 | Self-adaptive snapshot method for behavior of running red light |
| CN106887142A (en) * | 2017-04-17 | 2017-06-23 | 浙江宇视科技有限公司 | Changing method and device round the clock |
| CN107358803A (en) * | 2017-08-22 | 2017-11-17 | 哈尔滨理工大学 | A kind of traffic signal control system and its control method |
| CN109478233A (en) * | 2017-07-07 | 2019-03-15 | 丰田汽车欧洲股份有限公司 | The computer program for identifying method the reason of blocking in image sequence, executing the method, the computer readable recording medium comprising this computer program, the driving assistance system for being able to carry out the method |
| CN110097761A (en) * | 2019-05-17 | 2019-08-06 | 江苏航天大为科技股份有限公司 | Illumination adaptive video frequency vehicle flow detection mode switching system and method |
| CN115144547A (en) * | 2022-08-29 | 2022-10-04 | 四川九通智路科技有限公司 | Tunnel gas concentration detection method and device based on Internet of things and electronic equipment |
-
2007
- 2007-04-12 CN CNB2007100393782A patent/CN100504318C/en not_active Expired - Fee Related
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101727748A (en) * | 2009-11-30 | 2010-06-09 | 北京中星微电子有限公司 | Method, system and equipment for monitoring vehicles based on vehicle taillight detection |
| CN102044152A (en) * | 2010-11-19 | 2011-05-04 | 杭州海康威视系统技术有限公司 | Day and night video detecting method and device |
| CN102044152B (en) * | 2010-11-19 | 2013-10-30 | 杭州海康威视系统技术有限公司 | Day and night video detecting method and device |
| CN102842025A (en) * | 2011-06-25 | 2012-12-26 | 北京汉王智通科技有限公司 | Detection scene judgment method and detection scene judgment device of video image |
| CN102842025B (en) * | 2011-06-25 | 2016-01-06 | 北京汉王智通科技有限公司 | The detection scene determination methods of video image and device |
| CN102779272A (en) * | 2012-06-29 | 2012-11-14 | 惠州市德赛西威汽车电子有限公司 | Switching method for vehicle detection modes |
| CN102779272B (en) * | 2012-06-29 | 2016-01-06 | 惠州市德赛西威汽车电子股份有限公司 | A kind of changing method of vehicle detection mode |
| CN102881162A (en) * | 2012-09-29 | 2013-01-16 | 北京市交通信息中心 | Data processing and fusion method for large-scale traffic information |
| CN102881162B (en) * | 2012-09-29 | 2015-03-11 | 北京市交通信息中心 | Data processing and fusion method for large-scale traffic information |
| CN105187731B (en) * | 2015-09-18 | 2018-03-16 | 中电海康集团有限公司 | A kind of switching method round the clock and device based on preceding anti-collision warning |
| CN105187731A (en) * | 2015-09-18 | 2015-12-23 | 中电海康集团有限公司 | Day-night switching method and device based on forward collision warning |
| CN105389991A (en) * | 2015-12-03 | 2016-03-09 | 杭州中威电子股份有限公司 | Self-adaptive snapshot method for behavior of running red light |
| CN105389991B (en) * | 2015-12-03 | 2017-12-15 | 杭州中威电子股份有限公司 | A kind of adaptive Jaywalking snapshot method |
| CN106887142A (en) * | 2017-04-17 | 2017-06-23 | 浙江宇视科技有限公司 | Changing method and device round the clock |
| CN106887142B (en) * | 2017-04-17 | 2020-01-14 | 浙江宇视科技有限公司 | Day and night switching method and device |
| CN109478233A (en) * | 2017-07-07 | 2019-03-15 | 丰田汽车欧洲股份有限公司 | The computer program for identifying method the reason of blocking in image sequence, executing the method, the computer readable recording medium comprising this computer program, the driving assistance system for being able to carry out the method |
| CN109478233B (en) * | 2017-07-07 | 2023-08-08 | 丰田汽车欧洲股份有限公司 | Method for identifying cause of occlusion in image sequence, computer readable recording medium, driving assistance system |
| CN107358803A (en) * | 2017-08-22 | 2017-11-17 | 哈尔滨理工大学 | A kind of traffic signal control system and its control method |
| CN110097761A (en) * | 2019-05-17 | 2019-08-06 | 江苏航天大为科技股份有限公司 | Illumination adaptive video frequency vehicle flow detection mode switching system and method |
| CN115144547A (en) * | 2022-08-29 | 2022-10-04 | 四川九通智路科技有限公司 | Tunnel gas concentration detection method and device based on Internet of things and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100504318C (en) | 2009-06-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100504318C (en) | Switching method between videos for detecting vehicles during day-night | |
| CN104050620B (en) | Body detection device | |
| CN108986465B (en) | Method, system and terminal equipment for detecting traffic flow | |
| KR100459476B1 (en) | Apparatus and method for queue length of vehicle to measure | |
| RU2571368C1 (en) | Device for detecting three-dimensional objects, method of detecting three-dimensional objects | |
| CN103839415B (en) | Traffic flow based on pavement image feature identification and occupation rate information getting method | |
| CN104599502A (en) | Method for traffic flow statistics based on video monitoring | |
| CN103310006B (en) | A kind of area-of-interest exacting method in vehicle DAS (Driver Assistant System) | |
| CN109905847B (en) | Collaborative correction system and method for accumulated errors of GNSS blind area intelligent vehicle auxiliary positioning system | |
| CN102897086A (en) | Back car driving information detecting and prompting method and system | |
| CN105006154A (en) | Crossing traffic signal control optimization method in complex weather conditions | |
| CN110956146B (en) | Road background modeling method and device, electronic equipment and storage medium | |
| KR20100127074A (en) | System for controlling vehicle transport signal of intersection using day and night integrated traffic image detector | |
| CN102110366B (en) | Block-based accumulated expressway vehicle parking event detecting method | |
| CN106254827A (en) | A kind of group's mist Intelligent Recognition method for early warning and device thereof | |
| CN112133106B (en) | Geomagnetic characteristic-based traffic flow detection method and system | |
| CN102036016A (en) | Video detection method and system thereof | |
| CN103310199A (en) | Vehicle model identification method based on high-resolution remote sensing data | |
| CN110084218A (en) | The rainwater distributed data treating method and apparatus of vehicle | |
| KR20080004833A (en) | Apparatus and method for detecting a navigation vehicle in day and night according to luminous state | |
| CN105389991B (en) | A kind of adaptive Jaywalking snapshot method | |
| CN107554403A (en) | Based on prospect can identification car bulb autocontrol method and system | |
| CN104376316B (en) | License plate image acquisition method and device | |
| CN104782115A (en) | Image processing device for vehicle | |
| CN112365626A (en) | Vehicle-mounted camera picture processing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090624 Termination date: 20120412 |