CN105550663A - Cinema attendance statistical method and system - Google Patents

Abstract

The invention discloses a cinema attendance statistical method and system. The method comprises following steps: collecting a background image; extracting the edge of every seat region in the background image to generate a seat grid; obtaining monitoring images in at least one time period, calculating the differential image of every monitoring image in any one time period and the background image; accumulating all differential images in the time period to generate a target image in the time period; covering the seat grid on the target image to obtain the seat region of the target image; calculating the area of every seat region in the target image; with respect to the every seat region in the target image, obtaining the seating information of every seat in the time period based on the pixel distribution and seat region area of the seat region; and obtaining the attendance according to the seating information. According to the cinema attendance statistical method and system of the invention, the attendance of the cinema can be calculated accurately; therefore, judgment data are provided for obtaining real box office information.

Description

Theatre occupancy rate statistical method and system

Technical field

The present invention relates to technical field of image processing, particularly relate to a kind of theatre occupancy rate statistical method and system.

Background technology

The channel of film ticket sale has been enriched in the emergence of ecommerce, simultaneously also for the statistics of box office data adds new difficulty.For ensureing the correctness of box office data, supervision department has to send supervisor to carry out box office confirmation, but in the face of huge cinemas quantity, this mode is an utterly inadequate amount; Can directly obtain box office data by accessing each cinemas ticketing system, but this method can cause the significantly raising of administrative authority's communication system and software systems complexity, being also difficult to the imitation behavior identifying movie theatre simultaneously.

Therefore, need a kind of theatre occupancy rate statistical method and system badly, by attendance and then acquisition box office information, to solve the problems referred to above of prior art.

Summary of the invention

The invention provides a kind of theatre occupancy rate statistical method and system, utilize the video monitoring system being placed in the cinema projection Room, monitoring image is clearly obtained in rather dark Movie House, and the process passed through monitoring image and identification, the attendance of accurate calculating movie theatre, provides basis for estimation for obtaining real box office information.

One aspect of the present invention provides a kind of theatre occupancy rate statistical method, comprising: S1. gathers background image; Extract the edge of each seating area in background image, generate seat grid; S2. obtain the monitoring image of at least one time cycle, and calculate the difference image of every frame monitoring image and background image in a period of time in office; S3. difference images all in this time cycle is accumulated, generate the target image of this time cycle; Seat grid is covered on target image, obtains the seating area of target image; Calculate the area of each seating area in target image; S4. for each seating area in target image, based on pixel distribution and the seating area area of seating area, the information of taking one's seat at each seat within this time cycle is obtained; According to acquisition of information attendance of taking one's seat.

Preferably, gather background image and specifically comprise: gather m frame without image during spectators, and based on gather without image during spectators, according to formula 1 background extraction image;

$FB(i,j)=\frac{1}{m}\·\underset{p=1}{\overset{m}{\Σ}}{\mathrm{FOL}}_p(i,j)$ Formula 1

Wherein, m>1, and m ∈ N; The pixel value that FB (i, j) puts for (i, j) in background image; FOL _p(i, j) is the pixel value that p frame is put without (i, j) in image during spectators; P=1,2,3 ... m; I is positive integer, and j is positive integer.

Preferably, step S2 specifically comprises: the monitoring image obtaining at least one time cycle, and the difference image calculating n frame monitoring image and background image in a period of time in office according to formula 2;

BFD _k(i, j)=| F _k(i, j)-FB (i, j) | formula 2

Wherein, n>1, and n ∈ N; F _kthe pixel value that (i, j) puts for (i, j) in kth frame monitoring image in a period of time in office; BFD _k(i, j) is the pixel value that in kth frame difference image, (i, j) puts within this time cycle; K=1,2,3 ... n.

Preferably, step S2 also comprises: arrange with the null matrix of the picture element matrix homotype of monitoring image as initial accumulated matrix; Accumulate difference images all in this time cycle, the target image generating this time cycle specifically comprises:

S31. determine the segmentation threshold of every frame difference image respectively, and by segmentation threshold, split this difference image according to formula 3, generate the subdivision matrix corresponding with this difference image;

${\mathrm{BFS}}_k(i,j)=\left\{\begin{array}{cc}1& {\mathrm{BFD}}_k(i,j)\&GreaterEqual;{\mathrm{Th}}_k\\ -1& {\mathrm{BFD}}_k(i,j)<{\mathrm{Th}}_k\end{array}\right.$ Formula 3

S32. based on initial accumulated matrix and subdivision matrix corresponding to every frame difference image, accumulated image matrix is calculated according to formula 4;

${\mathrm{FSM}}_k(i,j)=min<max\{\&lsqb;{\mathrm{FSM}}_{k-1}(i,j)+{\mathrm{BFS}}_k(i,j)\&rsqb;,0\},255>$ Formula 4

S33. the binary-state threshold by presetting, carries out binary conversion treatment according to formula 5 pairs of accumulated image matrixes, generates the target image of this time cycle;

$FSE(i,j)=\left\{\begin{array}{cc}1& {\mathrm{FSM}}_n(i,j)\&GreaterEqual;th\\ 0& {\mathrm{FSM}}_n(i,j)<th\end{array}\right.$ Formula 5

Wherein, Th _kfor the segmentation threshold with kth frame difference image;

BFS _kthe pixel value that (i, j) puts for (i, j) in a kth subdivision matrix;

Max is max function, and min is minimum value function;

FSM _k-1(i, j) is the pixel value that in the accumulated matrix corresponding with kth-1 subdivision matrix, (i, j) puts;

As k=1, FSM _k-1(i, j)=FSM ₀(i, j), FSM ₀the pixel value that (i, j) puts for (i, j) in initial accumulated matrix;

FSM _k(i, j) is the pixel value that in the accumulated matrix corresponding with a kth subdivision matrix, (i, j) puts;

As k=n, FSM _k(i, j)=FSM _n(i, j), FSM _nthe pixel value that (i, j) puts for (i, j) in accumulated image matrix;

Th is default binary-state threshold;

The pixel value that in the target image that FSE (i, j) is this time cycle, (i, j) puts.

Preferably, determine that the segmentation threshold of every frame difference image specifically comprises respectively: the arithmetic mean and the standard deviation that calculate all pixel values in every frame difference image; For the pixel value being greater than arithmetic mean corresponding to this difference image and standard deviation sum in difference image, obtain the segmentation threshold of this difference image with maximum between-cluster variance criterion; Cover on target image by seat grid, the seating area obtaining target image specifically comprises: by the anchor point that coupling seat grid is corresponding with target image, cover on target image, obtain the seating area of target image by seat grid.

Preferably, step S33, after carrying out binary conversion treatment to accumulated image matrix, before the target image generating this time cycle, also comprises: for the accumulated image matrix after binary conversion treatment, adopt the first square templates to remove noise, adopt the second square templates filling cavity.

Preferably, for each seating area in target image, based on pixel distribution and the seating area area of seating area, the information of taking one's seat at each seat within this time cycle that obtains specifically comprises:

Calculate the bright spot area in the target image of this time cycle in each seating area, seat bright spot area in seating area being greater than area threshold is labeled as has people in this time cycle; Wherein, described bright spot to be pixel value be 1 point; Area threshold is the product of seating area area and default fixed coefficient;

According to taking one's seat, acquisition of information attendance specifically comprises: if what obtain is the monitoring image of a time cycle, have the seating capacity of people and total seating capacity, based on formula 6 calculated attendance rate according to being labeled as in target image in this time cycle;

$R=\frac{W_1}{W}$ Formula 6

Wherein, R is attendance, W ₁have the seating capacity of people for being labeled as in target image in this time cycle, W is total seating capacity.

Preferably, also comprise according to acquisition of information attendance of taking one's seat: if acquisition is respectively in period of time T ₁, T ₂t _ainterior monitoring image, then obtain attendance according to step S43, S44, S45:

S43. each seat is obtained respectively in period of time T ₁, T ₂t _ainterior information of taking one's seat;

S44. for each seat, statistics is labeled as the time cycle number b of people _x; If b _x>0.5a, then judge this use that loses one's seat;

S45. based on being judged to be occupied seating capacity and total seating capacity, by formula 7 calculated attendance rate;

$R=\frac{W_2}{W}$ Formula 7

Wherein, a is time cycle sum, a>1, and a ∈ N; T ₁=T ₂=T _a; X is seat number, x ∈ N; b _x<a, and b _x∈ N; W ₂for being judged to be occupied seating capacity.

The present invention also provides a kind of theatre occupancy rate statistical system, comprising: image acquisition units, seating area acquiring unit, image difference unit, Image accumulate unit and attendance computing unit; Wherein,

Image acquisition units, for gathering background image and the monitoring image at least one time cycle; And background image is sent to seating area acquiring unit, monitoring image is sent to image difference unit;

Image difference unit, for receiving monitoring image, calculating the difference image of every frame monitoring image and background image in a period of time in office, and difference image is sent to Image accumulate unit;

Image accumulate unit, for receiving difference image, carrying out difference images all in this time cycle accumulating the target image generating this time cycle, and target image being sent to seating area acquiring unit and attendance computing unit;

Seating area acquiring unit, for receiving background image, extracting the edge of each seating area in background image, generating seat grid; Seat grid is covered on target image, obtains the seating area of target image; Calculate the area of each seating area in target image; And the seating area of target image and seating area area are sent to attendance computing unit;

Attendance computing unit, for each seating area in target image, based on pixel distribution and the seating area area of seating area, obtains the information of taking one's seat at each seat within this time cycle; According to acquisition of information attendance of taking one's seat.

Preferably, Image accumulate unit comprises:

Difference image receiver module, for receiving difference image, and is sent to segmentation module by the difference image of reception;

Segmentation module, for being split difference image corresponding to this segmentation threshold by segmentation threshold, is generated the subdivision matrix corresponding with this difference image, and subdivision matrix is sent to accumulation module;

Accumulation module, for according to each subdivision matrix, calculates accumulated image matrix, and accumulated image matrix is sent to binarization block;

Binarization block, for carrying out binary conversion treatment to accumulated image matrix, and is sent to Morphological scale-space module by the accumulated image matrix after binary conversion treatment;

Morphological scale-space module, for for the accumulated image matrix after binary conversion treatment, the first square templates is adopted to remove noise, adopt the second square templates filling cavity, generate the target image of this time cycle, and target image is sent to seating area acquiring unit and attendance computing unit.

The present invention under the prerequisite not increasing administrative authority's communication system and software systems complexity, can utilize the video monitoring system being placed in the cinema projection Room, accurately calculating the attendance of movie theatre, providing basis for estimation for obtaining real box office information.

Accompanying drawing explanation

Fig. 1 is the first pass figure of theatre occupancy rate statistical method of the present invention;

Fig. 2 is the background image schematic diagram of theatre occupancy rate statistical method of the present invention and system;

Fig. 3 is the monitoring image schematic diagram of theatre occupancy rate statistical method of the present invention and system;

Fig. 4 is the difference image schematic diagram of theatre occupancy rate statistical method of the present invention and system;

Fig. 5 is the target image schematic diagram of theatre occupancy rate statistical method of the present invention and system;

Fig. 6 is the statistics schematic diagram of theatre occupancy rate statistical method of the present invention and system;

Fig. 7 is the second process flow diagram of theatre occupancy rate statistical method of the present invention;

Fig. 8 is multiple time cycle statistical flowsheet figure of theatre occupancy rate statistical method of the present invention;

Embodiment

For making object of the present invention, technical scheme and advantage clearly understand, enumerate preferred embodiment referring to accompanying drawing, the present invention is described in more detail.But it should be noted that, the many details listed in instructions are only used to make reader to have a thorough understanding, even if do not have these specific details also can realize these aspects of the present invention to one or more aspect of the present invention.

Existing box office information getting method exists that required supervisor is many, communication system and the problem such as software systems complexity is high.The invention provides a kind of theatre occupancy rate statistical method and system, utilize the video monitoring system being placed in the cinema projection Room, monitoring image is clearly obtained in rather dark Movie House, and the process passed through monitoring image and identification, the attendance of accurate calculating movie theatre, provides basis for estimation for obtaining real box office information.

One aspect of the present invention provides a kind of theatre occupancy rate statistical method, as shown in Figure 1, comprising:

S1. background image is gathered; Extract the edge of each seating area in background image, generate seat grid; Above-mentioned seating area comprises the border up and down at seat, and background image as shown in Figure 2.

In a preferred embodiment of the invention, gather background image and specifically comprise: gather m frame without image during spectators, and based on gather without image during spectators, according to formula 1 background extraction image;

$FB(i,j)=\frac{1}{m}\&CenterDot;\underset{p=1}{\overset{m}{\&Sigma;}}{\mathrm{FOL}}_p(i,j)$ Formula 1

Wherein, m>1, and m ∈ N; The pixel value that FB (i, j) puts for (i, j) in background image; FOL _p(i, j) is the pixel value that p frame is put without (i, j) in image during spectators; P=1,2,3 ... m; I is positive integer, and j is positive integer.

In a preferred embodiment of the invention, m gets 10.

Above-mentioned steps adopts the multi-frame mean method shown in formula 1 to carry out background modeling for m frame without image during spectators, obtains background image, said method can effectively stress release treatment on the impact of background image.

The edge extracting each seating area in background image need adopt existing pattern-recognition and edge extraction techniques, repeats no more herein.

S2. obtain the monitoring image of at least one time cycle, and calculate the difference image of every frame monitoring image and background image in a period of time in office.Monitoring image as shown in Figure 3.

In a preferred embodiment of the invention, step S2 specifically comprises:

Obtain the monitoring image of at least one time cycle, and calculate the difference image of n frame monitoring image and background image in a period of time in office according to formula 2.Difference image as shown in Figure 4;

BFD _k(i, j)=| F _k(i, j)-FB (i, j) | formula 2

Wherein, n>1, and n ∈ N; F _kthe pixel value that (i, j) puts for (i, j) in kth frame monitoring image; BFD _kthe pixel value that (i, j) puts for (i, j) in kth frame difference image; K=1,2,3 ... n.

In a preferred embodiment of the invention, monitoring image sum n=500; The monitoring image sum n in different time cycle can be identical, also can be different, determines as the case may be.

In a preferred embodiment of the invention, step S2 also comprises: arrange with the null matrix of the picture element matrix homotype of monitoring image as initial accumulated matrix, for the accumulation of difference image in step S3.

S3. difference images all in this time cycle is accumulated, generate the target image of this time cycle; Seat grid is covered on target image, obtains the seating area of target image; Calculate the area of each seating area in target image.

In this step, dividing processing is carried out to difference image and obtains motion target area, and the moving region image of each frame is accumulated, obtain target image.

In a preferred embodiment of the invention, accumulate difference images all in this time cycle, the target image generating this time cycle specifically comprises:

S31. determine the segmentation threshold of every frame difference image respectively, and by segmentation threshold, split difference image corresponding to this segmentation threshold according to formula 3, generate the subdivision matrix corresponding with this difference image;

${\mathrm{BFS}}_k(i,j)=\left\{\begin{array}{cc}1& {\mathrm{BFD}}_k(i,j)\&GreaterEqual;Th\left(k\right)\\ -1& {\mathrm{BFD}}_k(i,j)<Th\left(k\right)\end{array}\right.$ Formula 3

Wherein, Th _kfor the segmentation threshold corresponding with kth frame difference image; BFS _kthe pixel value that (i, j) puts for (i, j) in a kth subdivision matrix; High pixel region in subdivision matrix is moving region.

In a preferred embodiment of the invention, determine that the segmentation threshold of every frame difference image specifically comprises respectively: the arithmetic mean and the standard deviation that calculate all pixel values in every frame difference image; For the pixel value being greater than arithmetic mean corresponding to this difference image and standard deviation sum in difference image, obtain the segmentation threshold of this difference image with maximum between-cluster variance criterion.

In this step, segmentation threshold extracts in the pixel being greater than difference image pixel value arithmetic mean and standard deviation sum, can improve the accuracy of moving region segmentation, and adopt maximum between-cluster variance to choose segmentation threshold, misclassification probability can be made minimum.

Calculating and the maximum between-cluster variance criterion of described arithmetic mean, standard deviation are all known technological means, repeat no more herein.

S32. based on initial accumulated matrix and subdivision matrix corresponding to every frame difference image, accumulated image matrix is calculated according to formula 4;

${\mathrm{FSM}}_k(i,j)=min<max\{\&lsqb;{\mathrm{FSM}}_{k-1}(i,j)+{\mathrm{BFS}}_k(i,j)\&rsqb;,0\},255>$ Formula 4

Wherein, max is max function, obtains the maximal value in independent variable; Min is minimum value function, obtains the minimum value in independent variable.Operation result is each time limited in [0,255] scope by max function and min function by formula 4;

FSM _k-1(i, j) is and kth-1 subdivision matrix BFS _k-1the pixel value that in the accumulated matrix that (i, j) is corresponding, (i, j) puts;

As k=1, FSM _k-1(i, j)=FSM ₀(i, j), FSM ₀the pixel value that in the initial accumulated matrix that (i, j) is arranged for step 2, (i, j) puts;

FSM _k(i, j) is and subdivision matrix BFS _kthe pixel value that in the accumulated matrix that (i, j) is corresponding, (i, j) puts;

As k=n, FSM _k(i, j)=FSM _n(i, j), FSM _nthe pixel value that (i, j) puts for (i, j) in accumulated image matrix.

Step S32 achieves the accumulation to multi-frame difference image motion region, for the accurate calculating of attendance provides foundation.

S33. the binary-state threshold by presetting, carries out binary conversion treatment according to formula 5 pairs of accumulated image matrixes, generates the target image of this time cycle; Target image as shown in Figure 5.

$FSE(i,j)=\left\{\begin{array}{cc}1& {\mathrm{FSM}}_n(i,j)\&GreaterEqual;th\\ 0& {\mathrm{FSM}}_n(i,j)<th\end{array}\right.$ Formula 5

Wherein, th is default binary-state threshold, chooses as the case may be; The pixel value that in the target image that FSE (i, j) is this time cycle, (i, j) puts.

In a preferred embodiment of the invention, th=200.

In a preferred embodiment of the invention, step S33 is after carrying out binary conversion treatment to accumulated image matrix, before the target image generating this time cycle, also morphological operation will be carried out, comprise: for the accumulated image matrix after binary conversion treatment, the first square templates is adopted to carry out etching operation, to remove noise; The second square templates is adopted to carry out expansive working, with filling cavity.This step can obtain more complete target image, is conducive to improving later pixel statistical accuracy.

In a preferred embodiment of the invention, the first square templates is the square templates of 5 × 5, and the second square templates is the square templates of 7 × 7.

S34. seat grid is covered on target image, obtain the seating area of target image; Calculate the area of each seating area in target image.

Because background image and monitoring image are that the same video monitoring system being fixed on Movie House obtains, and target image is obtained by the difference to monitoring image, segmentation, accumulation, therefore the distribution of the seat of background image is identical with target image.Therefore, in step S34, the seat grid in background image is covered on target image with identical direction, the seating area of target image can be obtained.Calculate the area of each seating area in target image, each seating area in record object image and area thereof, as the foundation obtaining information of taking one's seat in step S4.

In a preferred embodiment of the invention, seat grid is covered on target image, the seating area obtaining target image specifically comprises: preset corresponding anchor point at seat grid and target image, by mating above-mentioned anchor point, seat grid is covered on target image, obtains the seating area of target image; Anchor point can be the summit of seat grid and target image, also can be other point.

S4. for each seating area in target image, based on pixel distribution and the seating area area of seating area, the information of taking one's seat at each seat within this time cycle is obtained; According to acquisition of information attendance of taking one's seat.Pixel distribution refers to the information of different pixels point occupied area and the ordered state of different pixels point.

In a preferred embodiment of the invention, step S4 specifically comprises:

S41. calculate the bright spot area in the target image of this time cycle in each seating area, seat bright spot area in seating area being greater than area threshold is labeled as has people in this time cycle; Wherein, bright spot to be pixel value be 1 point; Area threshold is the product of seating area area and default fixed coefficient.

Default fixed coefficient is chosen as the case may be, and in a preferred embodiment of the invention, default fixed coefficient is 0.2.

If what S42. obtain is the monitoring image of a time cycle, there are the seating capacity of people and total seating capacity, based on formula 6 calculated attendance rate according to being labeled as in target image in this time cycle;

$R=\frac{W_1}{W}$ Formula 6

Wherein, R is attendance, W ₁have the seating capacity of people for being labeled as in target image in this time cycle, W is total seating capacity.

In a preferred embodiment of the invention, also comprise according to acquisition of information attendance of taking one's seat:

If what obtain is respectively in period of time T ₁, T ₂t _ainterior monitoring image, then obtain attendance according to step S43, S44, S45.Wherein, a is time cycle sum, a>1, and a ∈ N; T ₁=T ₂=T _a.

S43. each seat is obtained respectively in period of time T ₁, T ₂t _ainterior flag state;

The time span of time cycle is chosen as the case may be, in a preferred embodiment of the invention, and T ₁=T ₂=T _a=10min.

Monitoring image due to the different time cycle is that the same video monitoring system being fixed on Movie House obtains, therefore the target image in different time cycle has identical seat distribution, can obtain each seat thus respectively in the information of taking one's seat in different time cycle.

S44. for each seat, statistics is labeled as the time cycle number b of people _x; If b _x>0.5a, then judge this use that loses one's seat; Wherein, x is seat number, x ∈ N; b _x<a, and b _x∈ N;

S45. based on being judged to be occupied seating capacity W ₂and total seating capacity, by formula 7 calculated attendance rate;

$R=\frac{W_2}{W}$ Formula 7

In this step, obtaining the omnidistance attendance of movie theatre, can provide basis for estimation for obtaining real box office information, omnidistance attendance statistics is as shown in Figure 6.

The present invention provides a kind of theatre occupancy rate statistical system on the other hand, comprising: image acquisition units, seating area acquiring unit, image difference unit, Image accumulate unit and attendance computing unit; Wherein,

Image acquisition units, for gathering background image and the monitoring image at least one time cycle; And background image is sent to seating area acquiring unit, monitoring image is sent to image difference unit;

Image difference unit, for receiving monitoring image, calculating the difference image of every frame monitoring image and background image in a period of time in office, and difference image is sent to Image accumulate unit;

Image accumulate unit, for receiving difference image, carrying out difference images all in this time cycle accumulating the target image generating this time cycle, and target image being sent to seating area acquiring unit and attendance computing unit;

Seating area acquiring unit, for receiving background image, extracting the edge of each seating area in background image, generating seat grid; Seat grid is covered on target image, obtains the seating area of target image; Calculate the area of each seating area in target image; And the seating area of target image and seating area area are sent to attendance computing unit;

Attendance computing unit, for each seating area in target image, based on pixel distribution and the seating area area of seating area, obtains the information of taking one's seat at each seat within this time cycle; According to acquisition of information attendance of taking one's seat.

In a preferred embodiment of the invention, image acquisition units is the video monitoring system being fixed on the cinema projection Room, adopts near-infrared light source light filling, and coordinates with narrow band pass filter, can obtain image and not by viewer senses clearly in rather dark movie theater.

In a preferred embodiment of the invention, Image accumulate unit comprises:

Difference image receiver module, for receiving difference image, and is sent to segmentation module by the difference image of reception;

Segmentation module, for being split difference image corresponding to this segmentation threshold by segmentation threshold, is generated the subdivision matrix corresponding with this difference image, and subdivision matrix is sent to accumulation module;

Accumulation module, for according to each subdivision matrix, calculates accumulated image matrix, and accumulated image matrix is sent to binarization block;

Binarization block, for carrying out binary conversion treatment to accumulated image matrix, and is sent to Morphological scale-space module by the accumulated image matrix after binary conversion treatment;

Morphological scale-space module, for for the accumulated image matrix after binary conversion treatment, the first square templates is adopted to remove noise, adopt the second square templates filling cavity, generate the target image of this time cycle, and target image is sent to seating area acquiring unit and attendance computing unit.

Fig. 7 is the second process flow diagram of theatre occupancy rate statistical method of the present invention, and shown in figure, described method comprises off-line part and online part.Off-line part comprises: the background modeling of background extraction image and seating area extract.Online part comprises image acquisition, background difference, Threshold segmentation, accumulative, the accumulative image binaryzation of difference, region area statistics and Output rusults.

Fig. 8 is the statistics schematic diagram of theatre occupancy rate statistical method of the present invention in multiple time cycle.Shown in figure, for each seat, add up in different time cycle T ₁, T ₂t _alabel information after carry out vote (namely in step S44, judging that whether seat occupied), draw statistics.

In a preferred embodiment of the invention, the hardware platform of image procossing adopts high-performance computer to realize, and software code adopts C language programming realization, and under MicrosoftVisualC++6.0 platform compilation run.

According to theatre occupancy rate statistical method of the present invention and system, accurately can calculating the attendance of movie theatre, providing basis for estimation for obtaining real box office information.

One of ordinary skill in the art will appreciate that all or part of step realized in above-described embodiment method is that the hardware that can carry out instruction relevant by program has come, this program can be stored in a computer read/write memory medium, as: ROM/RAM, magnetic disc, CD etc.

The above is only the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the premise without departing from the principles of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (10)

1. a theatre occupancy rate statistical method, is characterized in that, comprising:

S1. background image is gathered; Extract the edge of each seating area in background image, generate seat grid;

S2. obtain the monitoring image of at least one time cycle, and calculate the difference image of every frame monitoring image and background image in a period of time in office;

S3. difference images all in this time cycle is accumulated, generate the target image of this time cycle; Seat grid is covered on target image, obtains the seating area of target image; Calculate the area of each seating area in target image;

S4. for each seating area in target image, based on pixel distribution and the seating area area of seating area, the information of taking one's seat at each seat within this time cycle is obtained; According to acquisition of information attendance of taking one's seat.

2. the method for claim 1, is characterized in that, gathers background image and specifically comprises: gather m frame without image during spectators, and based on gather without image during spectators, according to formula 1 background extraction image;

$FB(i,j)=\frac{1}{m}\&CenterDot;\underset{p=1}{\overset{m}{\&Sigma;}}{\mathrm{FOL}}_p(i,j)$ Formula 1

Wherein, m>1, and m ∈ N; The pixel value that FB (i, j) puts for (i, j) in background image; FOL _p(i, j) is the pixel value that p frame is put without (i, j) in image during spectators; P=1,2,3 ... m; I is positive integer, and j is positive integer.

3. method as claimed in claim 2, it is characterized in that, step S2 specifically comprises: the monitoring image obtaining at least one time cycle, and the difference image calculating n frame monitoring image and background image in a period of time in office according to formula 2;

BFD _k(i, j)=| F _k(i, j)-FB (i, j) | formula 2

Wherein, n>1, and n ∈ N; F _kthe pixel value that (i, j) puts for (i, j) in kth frame monitoring image in a period of time in office; BFD _k(i, j) is the pixel value that in kth frame difference image, (i, j) puts within this time cycle; K=1,2,3 ... n.

4. method as claimed in claim 3, it is characterized in that, step S2 also comprises: arrange with the null matrix of the picture element matrix homotype of monitoring image as initial accumulated matrix;

Accumulate difference images all in this time cycle, the target image generating this time cycle specifically comprises:

S31. determine the segmentation threshold of every frame difference image respectively, and by segmentation threshold, split this difference image according to formula 3, generate the subdivision matrix corresponding with this difference image;

${\mathrm{BFS}}_k(i,j)=\left\{\begin{array}{cc}1& {\mathrm{BFD}}_k(i,j)\&GreaterEqual;{\mathrm{Th}}_k\\ -1& {\mathrm{BFD}}_k(i,j)<{\mathrm{Th}}_k\end{array}\right.$ Formula 3

S32. based on initial accumulated matrix and subdivision matrix corresponding to every frame difference image, accumulated image matrix is calculated according to formula 4;

FSM _k(i, j)=min<max{ [FSM _k-1(i, j)+BFS _k(i, j)], 0}, 255> formula 4

S33. the binary-state threshold by presetting, carries out binary conversion treatment according to formula 5 pairs of accumulated image matrixes, generates the target image of this time cycle;

$FSE(i,j)=\left\{\begin{array}{cc}1& {\mathrm{FSM}}_n(i,j)\&GreaterEqual;th\\ 0& {\mathrm{FSM}}_n(i,j)<th\end{array}\right.$ Formula 5

Wherein, Th _kfor the segmentation threshold with kth frame difference image;

BFS _kthe pixel value that (i, j) puts for (i, j) in a kth subdivision matrix;

Max is max function, and min is minimum value function;

FSM _k-1(i, j) is the pixel value that in the accumulated matrix corresponding with kth-1 subdivision matrix, (i, j) puts;

As k=1, FSM _k-1(i, j)=FSM ₀(i, j), FSM ₀the pixel value that (i, j) puts for (i, j) in initial accumulated matrix;

FSM _k(i, j) is the pixel value that in the accumulated matrix corresponding with a kth subdivision matrix, (i, j) puts;

As k=n, FSM _k(i, j)=FSM _n(i, j), FSM _nthe pixel value that (i, j) puts for (i, j) in accumulated image matrix;

Th is default binary-state threshold;

The pixel value that in the target image that FSE (i, j) is this time cycle, (i, j) puts.

5. method as claimed in claim 4, is characterized in that, determine that the segmentation threshold of every frame difference image specifically comprises respectively: the arithmetic mean and the standard deviation that calculate all pixel values in every frame difference image; For the pixel value being greater than arithmetic mean corresponding to this difference image and standard deviation sum in difference image, obtain the segmentation threshold of this difference image with maximum between-cluster variance criterion;

Cover on target image by seat grid, the seating area obtaining target image specifically comprises:

By the anchor point that coupling seat grid is corresponding with target image, seat grid is covered on target image, obtains the seating area of target image.

6. method as claimed in claim 5, is characterized in that, step S33, after carrying out binary conversion treatment to accumulated image matrix, before the target image generating this time cycle, also comprises:

For the accumulated image matrix after binary conversion treatment, adopt the first square templates to remove noise, adopt the second square templates filling cavity.

7. method as claimed in claim 6, is characterized in that,

For each seating area in target image, based on pixel distribution and the seating area area of seating area, the information of taking one's seat at each seat within this time cycle that obtains specifically comprises:

Calculate the bright spot area in the target image of this time cycle in each seating area, seat bright spot area in seating area being greater than area threshold is labeled as has people in this time cycle; Wherein, described bright spot to be pixel value be 1 point; Area threshold is the product of seating area area and default fixed coefficient;

According to taking one's seat, acquisition of information attendance specifically comprises:

If what obtain is the monitoring image of a time cycle, there are the seating capacity of people and total seating capacity, based on formula 6 calculated attendance rate according to being labeled as in target image in this time cycle;

$R=\frac{W_1}{W}$ Formula 6

Wherein, R is attendance, W ₁have the seating capacity of people for being labeled as in target image in this time cycle, W is total seating capacity.

8. method as claimed in claim 7, it is characterized in that, according to taking one's seat, acquisition of information attendance also comprises:

If what obtain is respectively in period of time T ₁, T ₂t _ainterior monitoring image, then obtain attendance according to step S43, S44, S45:

S43. each seat is obtained respectively in period of time T ₁, T ₂t _ainterior information of taking one's seat;

S44. for each seat, statistics is labeled as the time cycle number b of people _x; If b _x>0.5a, then judge this use that loses one's seat;

S45. based on being judged to be occupied seating capacity and total seating capacity, by formula 7 calculated attendance rate;

$R=\frac{W_2}{W}$ Formula 7

Wherein, a is time cycle sum, a>1, and a ∈ N; T ₁=T ₂=T _a; X is seat number, x ∈ N; b _x<a, and b _x∈ N; W ₂for being judged to be occupied seating capacity.

9. a theatre occupancy rate statistical system, is characterized in that, comprising: image acquisition units, seating area acquiring unit, image difference unit, Image accumulate unit and attendance computing unit; Wherein,

Image acquisition units, for gathering background image and the monitoring image at least one time cycle; And background image is sent to seating area acquiring unit, monitoring image is sent to image difference unit;

Image difference unit, for receiving monitoring image, calculating the difference image of every frame monitoring image and background image in a period of time in office, and difference image is sent to Image accumulate unit;

Image accumulate unit, for receiving difference image, carrying out difference images all in this time cycle accumulating the target image generating this time cycle, and target image being sent to seating area acquiring unit and attendance computing unit;

Seating area acquiring unit, for receiving background image, extracting the edge of each seating area in background image, generating seat grid; Seat grid is covered on target image, obtains the seating area of target image; Calculate the area of each seating area in target image; And the seating area of target image and seating area area are sent to attendance computing unit;

Attendance computing unit, for each seating area in target image, based on pixel distribution and the seating area area of seating area, obtains the information of taking one's seat at each seat within this time cycle; According to acquisition of information attendance of taking one's seat.

10. system as claimed in claim 9, it is characterized in that, Image accumulate unit comprises:

Difference image receiver module, for receiving difference image, and is sent to segmentation module by the difference image of reception;

Segmentation module, for being split difference image corresponding to this segmentation threshold by segmentation threshold, is generated the subdivision matrix corresponding with this difference image, and subdivision matrix is sent to accumulation module;

Accumulation module, for according to each subdivision matrix, calculates accumulated image matrix, and accumulated image matrix is sent to binarization block;

Binarization block, for carrying out binary conversion treatment to accumulated image matrix, and is sent to Morphological scale-space module by the accumulated image matrix after binary conversion treatment;

Morphological scale-space module, for for the accumulated image matrix after binary conversion treatment, the first square templates is adopted to remove noise, adopt the second square templates filling cavity, generate the target image of this time cycle, and target image is sent to seating area acquiring unit and attendance computing unit.