CN106127109A - A kind of pedestrian density computational methods of large-scale public place - Google Patents

Abstract

The invention discloses pedestrian density's computational methods of a kind of large-scale public place, pedestrian density's computational methods of this large-scale public place are based on the theory of Thiessen polygon, in a certain restriction region；With pedestrian's quantity and pedestrian's spacing as independent variable, the density using Thiessen polygon method to carry out pedestrian density under kinestate calculates, and improves the accuracy of calculating.

Description

A kind of pedestrian density computational methods of large-scale public place

Technical field

The invention belongs to urban rail transit planning field, the pedestrian density particularly relating to a kind of large-scale public place calculates Method.

Background technology

Pedestrian traffic basic data is the basis of research pedestrian traffic characteristic, owing to the behavior of pedestrian is complex, and its number Design and the optimization of pedestrian's infrastructure is directly influenced according to the accuracy processed, close for pedestrian in pedestrian's basic data processing The accurate process of degrees of data is then in the prevention internal crowded generation of pedestrian's large common facility, carry out in time pedestrains safety evacuation, The tool such as pedestrian's microscopic simulation is of great significance.

The research being related to pedestrian density's data acquisition being related to both at home and abroad in pedestrian flow data acquisition at present is also Achieving some accordingly to break through, the carrying out for the correlational study in later stage provides solid theory and practice basis.At this The research of aspect, external researchers are domestic relative to us, and the starting of research relatively early, is ground for density acquisition method Study carefully and be also carried out some substantial practical studies.In contrast, domestic this aspect research is mainly based on theoretical research. In initial stage, the method being all limited only to use original manual detection about the research of pedestrian density's data acquisition both at home and abroad Realize, while adding workload, also have ignored the certain district caused because of the change of the area of selected survey region Sharply increasing of flow of the people within territory and the congestion event that causes occurs；Recent decades mainly uses video acquisition method to pedestrian Traffic base data are acquired, but final or inevitably because we are traditional in density computational methods Cannot obtain a certain specific little for mobility and the factor such as observation interval, Continuous Observation number of times in region, pedestrian interval Pedestrian's crowding situation in region.Discreteness problem that pedestrian density's Data processing is faced with and pedestrian's distance variation problem All the time it is the key affecting pedestrian density's accuracy, as shown in Figure 3.

Summary of the invention

It is an object of the invention to provide pedestrian density's computational methods of a kind of large-scale public place, it is intended to solve traditional The not accurate enough problem of Static Density computational methods.

The present invention is achieved in that pedestrian density's computational methods reason with Thiessen polygon of a kind of large-scale public place Based on Lun, in a certain restriction region；With pedestrian's quantity and pedestrian's spacing as independent variable, use Thiessen polygon method to pedestrian Density carries out the density under kinestate and calculates.

Further, pedestrian density's computational methods of a kind of large-scale public place include:

Step one, collection dynamic row artificial abortion's video information；

Step 2, correction dynamic video Acquisition Error；

Step 3, determine the Thiessen polygon feature of static row artificial abortion；

Step 4, determine the conversion of map reference and actual coordinate；

Step 5, determine the algorithm of traditional static pedestrian density；

Step 6, determine the algorithm of dynamic density and combined density；

Step 7, carry out the analysis of two class pedestrian density's method data；

Step 8, pedestrian density's algorithm of assessment large-scale public place.

Further, described collection dynamic row artificial abortion's video information, select long a width of m × n's to limit region as pedestrian stream Video collection area, determines that the pedestrian entering restriction region within the video acquisition time all includes pedestrian density's meter in as basic point Calculate；For guaranteeing the dynamic menu visual angle collected and limiting region state close to the vertical shape, video acquisition place prioritizing selection limits The eminence of areas adjacent building；Requiring during video acquisition, shooting equipment is immovable and changes focal length, and camera lens is overlooked Angle requirement determines in advance and keeps constant.

Further, there is error in dynamic row artificial abortion's video of actual acquisition and the data demand needed for carrying out density calculating, The dynamic video error to gathering is needed to correct；Concrete error is as follows:

(1) camera lens depression angle error conversion, the video angle gathered in the ideal situation should be 90 with limiting region The right angle of degree, cancels the range difference brought due to vertical view, needs to add to take advantage of a coefficient of angularity y calculating density；

(2) near big and far smaller transparent effect Error processing, carries out " virtual to region on actual figure in the process to video Change " expand.

Further, determining the Thiessen polygon feature of static row artificial abortion, pedestrian density based on Thiessen polygon method calculates Method utilizes mapinfo software that each frame static state pedestrian's picture carries out Thiessen polygon dividing processing, and records the most each The area of Thiessen polygon, determines the Thiessen polygon feature that each moment static state pedestrian utilizes.

Further, determine the conversion of map reference and actual coordinate, in the actual length a width of m × n institute limited in region Use the two-dimensional coordinate system in units of rice, use in mapinfo software is to the processing procedure of static row artificial abortion Cartesian coordinate system in units of default-length, long a width of s × l, needed to calculate actual face before pedestrian density is calculated The long-pending ratio value k with area on map, formula (1) is as follows:

$k=\frac{s\×l}{m\×n}---\left(1\right).$

Further, determine the algorithm of traditional static pedestrian density, for there being the restriction region that measured area is A of N number of people, Standard defined formula is:

$Ds=\frac{N}{\left|A\right|}---\left(2\right)$

Further, determining the algorithm of dynamic density and combined density, limiting region area as A, pedestrian's quantity is the feelings of N Under condition, dynamic density algorithmic formula is as follows:

$D_{v^{\′}}=\frac{N}{{\mathrm{\Σ}}_{i=1}^N\left|A_i\right|}---\left(5\right)$

The formula of combined density algorithm is as follows:

$D_v=\frac{{\mathrm{\Σ}}_{i=1}^N\left|A_i\right|\·p\left(\stackrel{\→}{x}\right)}{i\·A}---\left(6\right).$

Further, carry out the analysis of two class pedestrian density's method data, use the side that data characteristics analysis and SPSS analyze Formula checks the accurate of pedestrian density's computational methods based on Thiessen polygon method, including:

(1) data characteristics analysis；

(2) SPSS one factor analysis of variance, the null hypothesis of one factor analysis of variance is: set H₀The average of=Static Density= The average of the average=combined density of dynamic density；

(3) SPSS paired sample T test is analyzed, and utilizes the paired sample T test of SPSS to analyze at three kinds of density algorithms Under the diversity of aggregate level that exists of the density value that draws.The null hypothesis of paired sample T test analysis is: set H₀=μ₁-μ₂ =0.μ₁And μ₂It is respectively first and second overall average.

Further, pedestrian density's algorithm of assessment large-scale public place, according to two class pedestrian density's computational methods data Analyze, three kinds of pedestrian density's computing formula are estimated.

The present invention is based on the theory of Thiessen polygon, in a certain restriction region；With pedestrian's quantity and pedestrian's spacing For independent variable, the density using Thiessen polygon method to carry out pedestrian density under kinestate calculates, and improves the accurate of calculating Property.

Accompanying drawing explanation

Fig. 1 is pedestrian density's computational methods flow chart of the large-scale public place that the embodiment of the present invention provides；

Fig. 2 is the Static Density calculating figure that the embodiment of the present invention provides；

Fig. 3 is the dynamic density calculating figure that the embodiment of the present invention provides；

Fig. 4 is the combined density calculating figure that the embodiment of the present invention provides.

Detailed description of the invention

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

Refer to Fig. 1 to Fig. 4:

Pedestrian density's computational methods of a kind of large-scale public place, based on the theory of Thiessen polygon, in a certain limit Determine in region；With pedestrian's quantity and pedestrian's spacing as independent variable, use Thiessen polygon method that pedestrian density is carried out kinestate Under density calculate.

Further, pedestrian density's computational methods of a kind of large-scale public place include:

S101, collection dynamic row artificial abortion's video information；

S102, correction dynamic video Acquisition Error；

S103, determine the Thiessen polygon feature of static row artificial abortion；

S104, determine the conversion of map reference and actual coordinate；

S105, determine the algorithm of traditional static pedestrian density；

S106, determine the algorithm of dynamic density and combined density；

S107, carry out the analysis of two class pedestrian density's method data；

S108, pedestrian density's algorithm of assessment large-scale public place.

Further, described collection dynamic row artificial abortion's video information, select long a width of m × n's to limit region as pedestrian stream Video collection area, determines that the pedestrian entering restriction region within the video acquisition time all includes pedestrian density's meter in as basic point Calculate；For guaranteeing the dynamic menu visual angle collected and limiting region state close to the vertical shape, video acquisition place prioritizing selection limits The eminence of areas adjacent building；Requiring during video acquisition, shooting equipment is immovable and changes focal length, and camera lens is overlooked Angle requirement determines in advance and keeps constant.

Further, there is error in dynamic row artificial abortion's video of actual acquisition and the data demand needed for carrying out density calculating, The dynamic video error to gathering is needed to correct；Concrete error is as follows:

(1) camera lens depression angle error conversion, the video angle gathered in the ideal situation should be 90 with limiting region The right angle of degree, cancels the range difference brought due to vertical view, needs to add to take advantage of a coefficient of angularity y calculating density；

(2) near big and far smaller transparent effect Error processing, carries out " virtual to region on actual figure in the process to video Change " expand.

Further, determining the Thiessen polygon feature of static row artificial abortion, pedestrian density based on Thiessen polygon method calculates Method utilizes mapinfo software that each frame static state pedestrian's picture carries out Thiessen polygon dividing processing, and records the most each The area of Thiessen polygon, determines the Thiessen polygon feature that each moment static state pedestrian utilizes.

Further, determine the conversion of map reference and actual coordinate, in the actual length a width of m × n institute limited in region Use the two-dimensional coordinate system in units of rice, use in mapinfo software is to the processing procedure of static row artificial abortion Cartesian coordinate system in units of default-length, long a width of s × l, needed to calculate actual face before pedestrian density is calculated The long-pending ratio value k with area on map, formula (1) is as follows:

$k=\frac{s\×l}{m\×n}---\left(1\right).$

Further, the algorithm of traditional static pedestrian density is determined；

What the normal process of measurement conventional densities directly reflected is the definition to Static Density, and it is specifically defined is to measure Area is the pedestrian's quantity limited on region in each unit are of A, such as Fig. 2.

For there being the restriction region that measured area is A of N number of people, standard defined formula is:

$Ds=\frac{N}{\left|A\right|}---\left(2\right)$

Further, the algorithm of dynamic density and combined density is determined；

Thiessen polygon method is different from traditional Static Density computational methods, and its core is every by limit in the A of region One pedestrian regards a base station as, and adjacent two base stations are made perpendicular bisector and constituted Thiessen polygon A one by one_i.According to the method, The density computing formula of each little Thiessen polygon is:

$p_i\left(\stackrel{\→}{x}\right)=\frac{1}{\left|A_i\right|}---\left(3\right)$

Overall formula is:

$p_i\left(\stackrel{\→}{x}\right)=\left\{\begin{array}{c}\frac{1}{\left|A_i\right|}:\stackrel{\→}{x}\∈A_i\\ 0:otherwise\end{array}\right\}andp\left(\stackrel{\‾}{x}\right)=\underset{i}{\Σ}{p}_i\left(\stackrel{\→}{x}\right)---\left(4\right)$

Being the density to all Thiessen polygon to sue for peace, the restriction region A selected yet with us leads to It is often the rectangle of rule, and the restriction region A selected by Thiessen polygon method_iIt is typically irregular polygon, simultaneously ∑_iA_i (the Thiessen polygon area sum that the most any moment the same area is produced by Thiessen polygon method is quiet less than or equal to tradition for≤A Restriction region area in density of states algorithm)；

Therefore, limiting region area as A, in the case of pedestrian's quantity is N, dynamic density algorithm (such as Fig. 3) formula is such as Under:

$D_{v^{\′}}=\frac{N}{{\mathrm{\Σ}}_{i=1}^N\left|A_i\right|}---\left(5\right)$

The formula of combined density algorithm (such as Fig. 4) is as follows:

$D_v=\frac{{\mathrm{\Σ}}_{i=1}^N\left|A_i\right|\·p\left(\stackrel{\→}{x}\right)}{i\·A}---\left(6\right)$

Now as a example by passage student at the bottom of No. four teaching building of Ningbo Engineering College goes to school and leaves school pedestrian stream, carry out application demonstration.

1) dynamic row artificial abortion's video information is gathered

Select long a width of 27m × 3m's to limit region as pedestrian stream video collection area, determine within the video acquisition time The pedestrian entering restriction region all includes pedestrian density's calculating in as basic point.For guaranteeing the dynamic menu visual angle and the restriction that collect Region state close to the vertical shape, video acquisition place prioritizing selection limits the eminence of areas adjacent building.Require at video acquisition During, shooting equipment is immovable and changes focal length, and camera lens depression angle requires determine in advance and keep constant.

2) dynamic video Acquisition Error is corrected

Owing to dynamic row artificial abortion's video of actual acquisition and the data demand needed for carrying out density calculating exist certain error, So that the dynamic video error gathered is corrected.Concrete error is as follows:

1. camera lens depression angle error conversion.The video angle gathered in the ideal situation should be 90 degree with limiting region Right angle such that it is able to cancel due to the range difference that brings of vertical view, but owing to actual situation cannot accomplish ideal effect, Calculating density to need to add to take advantage of a coefficient of angularity y, the value in calculating angle of depression value and distance y afterwards is 0.72, but calculates in density Method relatively in add and take advantage of identical coefficient that result does not produce impact, therefore can ignore.

The most near big and far smaller transparent effect Error processing.Due to gather video information in limit region show be not advise Rectangle then, therefore needs in the process to video that region on actual figure is carried out " virtualization " and expands.

3) the Thiessen polygon feature of static row artificial abortion is determined

Pedestrian density's computational methods based on Thiessen polygon method need to utilize mapinfo software to each frame static state pedestrian Picture carries out Thiessen polygon dividing processing, and the area of the most each Thiessen polygon of record, so that it is determined that each moment is quiet The Thiessen polygon feature that state pedestrian utilizes.

4) conversion of map reference and actual coordinate is determined

The two-dimensional coordinate system in units of rice is used at the actual a width of 27m × 3m of length limited in region, and The cartesian coordinate system in units of default-length is used in mapinfo software is to the processing procedure of static row artificial abortion, Long a width of 880 × 430, before pedestrian density is calculated, therefore need the ratio value k calculating real area with area on map, public Formula (1) is as follows:

$k=\frac{s\×l}{m\×n}=\frac{880\×430}{27\×3}=4671---\left(1\right)$

5) algorithm of traditional static pedestrian density is determined

What the normal process of measurement conventional densities directly reflected is the definition to Static Density, and it is specifically defined is to measure Area is the pedestrian's quantity limited on region in each unit are of A.

For have the measured area of N number of people be A limit region (according to calculating real area as 81m^-2), standard definition public affairs Formula is:

$Ds=\frac{N}{\left|A\right|}=\frac{N}{81}---\left(2\right)$

6) algorithm of dynamic density and combined density is determined.

Thiessen polygon method is different from traditional Static Density computational methods, and its core is every by limit in the A of region One pedestrian regards a base station as, and adjacent two base stations are made perpendicular bisector and constituted Thiessen polygon A one by one_i.According to the method, The density computing formula of each little Thiessen polygon is:

$p_i\left(\stackrel{\→}{x}\right)=\frac{1}{\left|A_i\right|}---\left(3\right)$

Overall formula is:

$p_i\left(\stackrel{\→}{x}\right)=\left\{\begin{array}{c}\frac{1}{\left|A_i\right|}:\stackrel{\→}{x}\∈A_i\\ 0:otherwise\end{array}\right\}andp\left(\stackrel{\‾}{x}\right)=\underset{i}{\Σ}{p}_i\left(\stackrel{\→}{x}\right)---\left(4\right)$

Being the density to all Thiessen polygon to sue for peace, the restriction region A selected yet with us leads to It is often the rectangle of rule, and the restriction region A selected by Thiessen polygon method_iIt is typically irregular polygon, simultaneously ∑_iA_i (the Thiessen polygon area sum that the most any moment the same area is produced by Thiessen polygon method is quiet less than or equal to tradition for≤A Restriction region area in density of states algorithm).

Therefore, limiting region area as A, in the case of pedestrian's quantity is N, dynamic density algorithmic formula is as follows:

$D_{v^{\′}}=\frac{N}{{\mathrm{\Σ}}_{i=1}^N\left|A_i\right|}---\left(5\right)$

The algorithmic formula of combined density is as follows:

$D_v=\frac{{\mathrm{\Σ}}_{i=1}^N\left|A_i\right|\·p\left(\stackrel{\→}{x}\right)}{i\·A}---\left(6\right)$

The density value such as table 1-table 2 that three kinds of density computational methods that student goes to school and leaves school obtain can be obtained according to algorithm above Shown in.

The density value of table 1 large-scale public place densimeter algorithm

Place: campus；Time: the morning 7:00～7:15

The density value of table 2 large-scale public place densimeter algorithm

Place: campus；Time: noon 11:40～11:55

7) analysis of two class pedestrian density's method data is carried out

The mode using data characteristics analysis and SPSS to analyze checks the pedestrian density side of calculating based on Thiessen polygon method The accuracy of method.

(1) data characteristics is analyzed as shown in table 3:

Table 3 large-scale public place pedestrian density's characteristic table

Place: campus；Time: the morning 7:00～7；15

Table 4 large-scale public place pedestrian density's characteristic table

Place: campus；Time: noon 11:40～11:55

(2) SPSS one factor analysis of variance

The null hypothesis of one factor analysis of variance is: set H₀The average of the average=dynamic density of=Static Density=the closeest The average of degree, i.e. control variable (density value size) average of (three kinds of different density algorithms) observational variable under varying level Without notable change (i.e. control effect is 0), it is meant that observational variable is not produced notable by the change of control variable varying level Impact, as shown in table 4～table 5.

Table 5 homogeneity test of variance

Density size

Table 6 ANOVA

Density size

(3) SPSS paired sample T test is analyzed

It is overall that the paired sample T test utilizing SPSS analyzes that the density value that draws under three kinds of density algorithms exists The diversity of level.The null hypothesis of paired sample T test analysis is: set H₀=μ 1-μ₂=0.μ 1 and μ₂Be respectively first and Second overall average.

The paired sample T test analysis result of student's density value of upper class hour is as shown in table 7～table 9:

Table 7 paired samples statistic

Table 8 paired samples correlation coefficient

Table 9 paired samples T checks

The paired sample T test analysis result of student's density value of lower class hour is as shown in table 10～table 12:

Table 10 paired samples statistic

Table 11 paired samples correlation coefficient

Table 12 paired samples T checks

8) pedestrian density's algorithm based on Thiessen polygon method is assessed

(1) for pedestrian density's situation analysis of student's upper class hour:

1. showing from variance and standard deviation and see Static Density dynamic density to be significantly less than fluctuating margin, main cause exists Only affected by pedestrian's quantity and less in the fluctuation of the situation descending people quantity of little density in Static Density, and dynamic density By the computing of Thiessen polygon method make to be expert at people's quantity identical with Static Density in the case of add the spacing of pedestrian Impact.

2. in the size of density value, Static Density to be also significantly less than dynamic density, and this is that dynamic density existence " is taken advantage of Deceive effect ", its reason is the gross area ∑ that the algorithm of dynamic density does not accounts for Thiessen polygon_iA_iAnd limit between the A of region Relation, the gross area ∑ that obtained by Thiessen polygon method in the case of little passenger flow is assembled_iA_iTo be much smaller than and to limit region A, Therefore, dynamic density is bigger than normal than real density in the case of little passenger flow is assembled.

By the T check analysis in the case of the pedestrian density of class hour upper to student, can be concluded that

3. table 6 gives the descriptive statistic amount of the sample average inspection that three kinds of density samples are mutually paired；Table 7 gives Dependency between three kinds of density samples, wherein dynamic density and Static Density, the correlation coefficient of combined density be respectively- 0.083 and-0.178, illustrate that the data characteristics of dynamic density in the case of little passenger flow is assembled and result have a tremendous difference really, And the correlation coefficient of Static Density and combined density reaches 0.377, illustrate to have each other obvious dependency on data result But not up to significant correlation, reason is consistent with what " deception effect " was analyzed；

4. table 8 is the results list of paired sample inspection, and double tail P values of three kinds of density samples are 0.000, it is believed that exist Two kinds of probabilities: there is not dependency between three kinds of density datas: or exist between three kinds of density is nonlinear relevant. From the point of view of the interpretation of result of table 7 and the source data of density are identical situation, it should be non-owing to also existing between three kinds of density Linearly relevant is caused.Combined density algorithm Static Density to be substantially better than and dynamic in the case of student attends class as can be seen here Density of states algorithm.

(2) for pedestrian density's situation analysis of student's lower class hour:

1. show from variance and standard deviation and see that three kinds of density change over broken line graph and have in fluctuating margin and density values Certain difference.Static Density and the variance of dynamic density and standard deviation show that the fluctuating margin of two kinds of density is more or less the same, main Reason is wanted to be restriction region A and the Thiessen polygon gross area ∑ of dynamic density of Static Density in this case_iA_iDifference is not Greatly, illustrate that in the case of the pedestrian density of class hour, the effect of Thiessen polygon method is inconspicuous under student.

2. fluctuating margin and the density values of combined density is all slightly larger than Static Density and dynamic density (mainly formula 6 The impact that the weight of Midst density produces).

By the T check analysis in the case of the pedestrian density to student's lower class hour, it may be concluded that

3. table 9 gives the descriptive statistic amount of the sample average inspection that three kinds of density samples are mutually paired；Table 10 is the most anti- The dependency between three kinds of density samples, Static Density and dynamic density, correlation coefficient 0.693 and of combined density are answered 0.778, both of which is more than significance level 0.05, height correlation each other be described in data, has all reached significance level, demonstrate,proves Bright student finish class in the case of the data of Static Density and the data of dynamic density, combined density be more or less the same, and the closeest Spending the correlation coefficient between combined density is 0.492 slightly below significance level 0.05, illustrates between the two due to algorithm Difference and cause degree of association lower slightly；

4. table 11 is the results list of paired sample inspection, and double tail P values of three kinds of density samples are 0.000, and reason is same On, owing to there is nonlinear relevant cause between three kinds of density.Therefore, in the case of student finishes class, combined density Algorithm slightly advantage is the most little with the difference of Static Density and dynamic density.

(3) one factor analysis of variance assessment

As shown in table 4, homogeneity test of variance and the significance that obtains is 0.06, the significance level arranged more than system 0.05, therefore can be with rejection null hypothesis, the average situation of three kinds of density has significant difference；And according to single factor test variance table 5 it will be seen that significance P value is 0.000 to be less than significance level 0.05, and therefore rejection null hypothesis also concludes that three kinds The algorithm of different densities is significant on the impact of density value size.As can be seen here the algorithm of three kinds of pedestrian densities be also exist bright Significant difference is different, even illustrating still to there is difference between three under similar circumstances.

In sum, it can be deduced that conclusion:

(1) dynamic density and combined density are owing to have employed Thiessen polygon method, add between pedestrian than Static Density The variables of distance so that pedestrian density is worth closer to reality.

(2) under different passenger flow gathering situations, the applicable situation of Static Density, dynamic density and combined density is to differ Sample:

Similar student attend class such little passenger flow assemble situation under, combined density than Static Density closer to actual shape Condition, and dynamic density there will be the situation of distortion, there is bigger error.

Finishing class under the situation that such large passenger flow assembles similar student, three kinds of mean density value difference is smaller, undulating Condition there are differences, and dynamic density and combined density are better than Static Density on practical situation, but dynamic density and the closeest Suitable situation between degree is more or less the same.

Utilize technical solutions according to the invention, or those skilled in the art be under the inspiration of technical solution of the present invention, Design similar technical scheme, and reach above-mentioned technique effect, all fall into protection scope of the present invention.

Claims (10)

1. pedestrian density's computational methods of a large-scale public place, it is characterised in that the pedestrian of described large-scale public place Density computational methods are based on the theory of Thiessen polygon, in a certain restriction region；With pedestrian's quantity and pedestrian's spacing it is Independent variable, the density using Thiessen polygon method to carry out pedestrian density under kinestate calculates.

2. pedestrian density's computational methods of large-scale public place as claimed in claim 1, it is characterised in that a kind of large-scale public Pedestrian density's computational methods in place include:

Step one, collection dynamic row artificial abortion's video information；

Step 2, correction dynamic video Acquisition Error；

Step 3, determine the Thiessen polygon feature of static row artificial abortion；

Step 4, determine the conversion of map reference and actual coordinate；

Step 5, determine the algorithm of traditional static pedestrian density；

Step 6, determine the algorithm of dynamic density and combined density；

Step 7, carry out the analysis of two class pedestrian density's method data；

Step 8, pedestrian density's algorithm of assessment large-scale public place.

3. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that described collection is moved State pedestrian stream video information, selects long a width of m × n's to limit region as pedestrian stream video collection area, determines at video acquisition The pedestrian entering restriction region in time all includes pedestrian density's calculating in as basic point；For guaranteeing the dynamic menu visual angle collected With restriction region state close to the vertical shape, video acquisition place prioritizing selection limits the eminence of areas adjacent building；Require regarding Frequently in gatherer process, shooting equipment is immovable and changes focal length, and camera lens depression angle requires determine in advance and keep constant.

4. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that moving of actual acquisition There is error in state pedestrian stream video and the data demand needed for carrying out density calculating, needs the dynamic video error to gathering to carry out Correct；Concrete error is as follows:

(1) camera lens depression angle error conversion, the video angle gathered in the ideal situation should be 90 degree with limiting region Right angle, cancels the range difference brought due to vertical view, needs to add to take advantage of a coefficient of angularity y calculating density；

(2) near big and far smaller transparent effect Error processing, carries out virtualization to region on actual figure in the process to video and expands Greatly.

5. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that determine static pedestrian The Thiessen polygon feature of stream, pedestrian density's computational methods based on Thiessen polygon method utilize mapinfo software to each frame Static pedestrian's picture carries out Thiessen polygon dividing processing, and the area of the most each Thiessen polygon of record, determines per a period of time Carve the Thiessen polygon feature that static pedestrian utilizes.

6. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that determine map reference With the conversion of actual coordinate, used the two-dimensional coordinate in units of rice at the actual a width of m × n of length limited in region System, uses the cartesian coordinate in units of default-length in mapinfo software is to the processing procedure of static row artificial abortion System, long a width of s × l, before pedestrian density is calculated, need the ratio value k calculating real area with area on map, formula (1) As follows:

$k=\frac{s\×l}{m\×n}---\left(1\right).$

7. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that determine traditional static The algorithm of pedestrian density, for there being the restriction region that measured area is A of N number of people, standard defined formula is:

$Ds=\frac{N}{\left|A\right|}---\left(2\right).$

8. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that determine dynamic density With the algorithm of combined density, limiting region area as A, in the case of pedestrian's quantity is N, dynamic density algorithmic formula is as follows:

$D_{v^{\′}}=\frac{N}{{\Σ}_{i=1}^N\left|A_i\right|}---\left(5\right)$

The formula of combined density algorithm is as follows:

$D_v=\frac{{\Σ}_{i=1}^N\left|A_i\right|\·p\left(\stackrel{\→}{x}\right)}{i\·A}---\left(6\right).$

9. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that carry out two class pedestrians The analysis of density method data, the mode using data characteristics analysis and SPSS to analyze checks pedestrian based on Thiessen polygon method Density computational methods accurate, including:

(1) data characteristics analysis；

(2) SPSS one factor analysis of variance, the null hypothesis of one factor analysis of variance is: set H₀The average of=Static Density=dynamically The average of the average=combined density of density；

(3) SPSS paired sample T test is analyzed, and utilizes the paired sample T test of SPSS to analyze under three kinds of density algorithms The diversity of the aggregate level that the density value drawn exists.The null hypothesis of paired sample T test analysis is: set H₀=μ₁-μ₂= 0。μ₁And μ₂It is respectively first and second overall average.

10. pedestrian density's computational methods of large-scale public place as claimed in claim 2, it is characterised in that assess large-scale public affairs Three kinds of pedestrian densities, according to the analysis of two class pedestrian density's computational methods data, are calculated public affairs by pedestrian density's algorithm in place altogether Formula is estimated.