CN102156099A - Method and system for detecting atmospheric pollutants - Google Patents

Abstract

The invention is suitable for the field of environmental protection and provides a method and system for detecting atmospheric pollutants. The method comprises the following steps of: collecting an image of the atmospheric pollutants, and converting the collected video image to a gray level image; comparing a preselected reference region with a selected target region in the gray level image to determine whether the target region is a smoke plume region; if so, compensating brightness of the smoke plume region according to a relation of the reference region and the smoke plume region; carrying out gray level histogram statistics on the compensated smoke plume region, and calculating a gravity center of a gray level histogram; and figuring a blackness value of the smoke plume region according to the gravity center of the gray level histogram. The technical scheme, provided by the invention, overcomes the defect of subjectivity due to artificial detection and has the advantages of improving blackness detection.

Description

A kind of detection method of atmosphere pollution and system

Technical field

The invention belongs to the environmental protection field, relate in particular to a kind of detection method and system of atmosphere pollution.

Background technology

Along with development of science and technology, pollutant in the atmosphere is more and more, the detection method of existing atmosphere pollution specifically can comprise: the flue gas of artificially observing chimney discharge, flue gas and Ringelmen smoke chart that chimney is discharged contrast to determine the blackness of exhaustion value of series.Wherein Ringelmen smoke chart can be referring to " emission standard of air pollutants for boilers " GB1327-91.

According to the technical scheme that prior art provided, find to exist in the prior art following technical matters:

The method that prior art provides is owing to be artificial observation, and the human factor influence is big, and artificial observed efficiency is low.

Summary of the invention

The purpose of the embodiment of the invention is to provide a kind of detection method of atmosphere pollution, and it is big to be intended to solve in the method for the prior art the human factor influence, inefficient problem.

The embodiment of the invention is achieved in that the detection method that the invention provides a kind of atmosphere pollution, and described method comprises the steps:

Gather the image of atmosphere pollution, convert the video image that collects to gray level image;

Selected target area in more previously selected reference zone and the gray level image determines whether this target area is the plume zone;

As after being defined as the plume zone,, the plume zone is carried out the compensation of brightness according to the relation in reference zone and plume zone;

Statistics of histogram is carried out in plume zone after the compensation, calculate the center of gravity of grey level histogram;

Draw the blackness value in plume zone according to the center of gravity calculation of grey level histogram.

The present invention also provides a kind of detection system of atmosphere pollution, and described system comprises:

The image acquisition converting unit is used to gather the image of atmosphere pollution, converts the video image that collects to gray level image;

Comparing unit, the selected target area that is used for more previously selected reference zone and gray level image determines whether this target area is the plume zone;

Compensating unit is used for after being defined as the plume zone, according to the relation in reference zone and plume zone, the plume zone is carried out the compensation of brightness;

The statistical computation unit is used for statistics of histogram is carried out in the plume zone after the compensation, calculates the center of gravity of grey level histogram;

Computing unit is used for drawing according to the center of gravity calculation of grey level histogram the blackness value in plume zone.

The embodiment of the invention compared with prior art, beneficial effect is: technical scheme of the present invention is owing to be that the image of gathering is operated the calculating of finishing blackness value accordingly, realize the intellectuality that blackness detects, overcome the defective of the subjectivity of artificial detection, improve the advantage that blackness detects.

Description of drawings

Fig. 1 is the process flow diagram of the detection method of atmosphere pollution provided by the invention;

Fig. 2 is the smog region decision process flow diagram that the embodiment of the invention provides;

Fig. 3 is the target area compensation process flow diagram that the embodiment of the invention provides;

Fig. 4 is the process flow diagram that image carries out statistics of histogram after the compensation that provides of the embodiment of the invention;

Fig. 5 is the structural drawing of the detection system of atmosphere pollution provided by the invention.

Embodiment

In order to make purpose of the present invention, technical scheme and advantage clearer,, the present invention is further elaborated below in conjunction with drawings and Examples.Should be appreciated that specific embodiment described herein only in order to explanation the present invention, and be not used in qualification the present invention.

The invention provides a kind of detection method of atmosphere pollution, this method specifically comprises the steps: as shown in Figure 1

The image of S11, collection atmosphere pollution converts the video image that collects to gray level image;

Selected target area in S12, more previously selected reference zone and the gray level image determines whether this target area is the plume zone; As be the plume zone, then carry out subsequent step (S13-S15), otherwise the output blackness value is 0.

S13, according to the relation in reference zone and plume zone, the plume zone is carried out the compensation of brightness;

S14, statistics of histogram is carried out in the plume zone after the compensation, calculate the center of gravity of grey level histogram;

S15, draw the blackness value in plume zone according to the center of gravity calculation of grey level histogram.

Need to prove that the method for realization S12 can comprise specifically as shown in Figure 2:

S121, set up the background model of reference zone in advance, need to prove the above-mentioned essentially identical zone of gray scale that preferentially is chosen in the non smoke time domain at reference zone.The method of setting up background model as, single Gaussian distribution is obeyed in the distribution of reference zone pixel, the Gauss model that can set up reference zone by the average and the variance of pixel by the training of a period of time, obtains stable reference background model.Specifically set up mode can for:

$\mathrm{\μ}(i,j)=\frac{1}{N}(I_1(i,j)+I_2(i,j)+...+I_N(i,j));---\left(1\right)$

$\mathrm{\σ}(i,j)=\frac{1}{N}((I_1(i,j)-\mathrm{\μ}(i,j))+(I_2(i,j)-\mathrm{\μ}(i,j))+...+(I_N(i,j)-\mathrm{\μ}(i,j)));---\left(2\right)$

Wherein, (i, j) the capable j of the i of presentation video is listed as I respectively _N((i j) is the average of reference zone corresponding pixel points to μ, and (i j) is the variance of reference zone corresponding pixel points to σ for i, the j) pixel value of the reference zone corresponding pixel points of expression N frame.

S122, according to the pixel of target area (being plume zone to be detected) and the background model of reference zone, determine that those pixels belong to the smog pixel in the target area.

The concrete grammar of its realization can for, each pixel for the target area of every two field picture utilizes difference image and variance to compare and judges whether current point is the smog point, concrete determination methods is:

$D(i,j)=\left\{\begin{array}{c}255,\mathrm{abs}(\mathrm{lr}(i,j)-\mathrm{\&mu;}(i,j))>=N_1\&CenterDot;\mathrm{\&sigma;}(i,j)\\ 0,\mathrm{abs}(\mathrm{lr}(i,j)-\mathrm{\&mu;}(i,j))<N_1\&CenterDot;\mathrm{\&sigma;}(i,j)\end{array}\right.---\left(3\right)$

Wherein, (i j) is the pixel value of target area corresponding pixel points to lr; (i j) is value after the binaryzation to D, and (i, j)=255 o'clock, this pixel is the smog pixel, otherwise is non-smog pixel as D.

S123, surpass threshold value, determine that then this target area is the plume zone, otherwise be non-plume zone as the smog pixel in the target area.

Need to prove that above-mentioned threshold value can be concrete numerical value, for example 100,1000 etc., can certainly be the smog pixel and the ratio of total pixel, for example 50%, 60% or the like.

Need to prove, the implementation method of above-mentioned S13 specifically as shown in Figure 3, can for:

Mean flow rate in S131, the calculating setting-up time internal reference zone.

Concrete computing method can for:

$\mathrm{Lr}=\frac{1}{M*N*n}\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(I_1(i,j)+...I_n(i,j))---\left(4\right)$

Wherein, M is the reference zone width, and N is the reference zone height, and n represents to add up the reference zone in the n frame; Lr is the mean flow rate of reference zone.

S132, according to the mean flow rate of reference zone and the brightness relationship in plume zone, luminance compensation is carried out in the plume zone.

Concrete compensation method is:

Ld(i，j)＝lr(i，j)-Lr (5)

Wherein, (i is to the pixel value behind the plume regional compensation j) to Ld.

Optionally, the method that realizes S14 specifically as shown in Figure 4, can for:

S141, the plume zone after will compensating are divided into the piece of 3*3;

S142, each image subblock is carried out the grey level histogram hist that gray-scale statistical obtains each height piece correspondence ₁[n] ..., hist ₉[n]; Wherein the span of n is 0-255;

S143, calculate the center of gravity of the grey level histogram of each sub-piece;

Concrete computing method can for:

$\mathrm{hist}\_x_i=\frac{\underset{n=0}{\overset{255}{\mathrm{\&Sigma;}}}n*\mathrm{his}{t}_i\left[n\right]}{\underset{n=0}{\overset{255}{\mathrm{\&Sigma;}}}{\mathrm{hist}}_i\left[n\right]}---\left(6\right)$

Hist wherein _iThe grey level histogram of i sub-piece of [n] expression; Hist_x _iThe grey level histogram center of gravity of representing i sub-piece.

Optionally, the step that realizes S15 is specifically as follows: (support vector machine SVM) calculates the blackness level as the good support vector machine of input vector input training in advance with the center of gravity of each sub-piece of grey level histogram.

Wherein, this blackness utmost point adopts the general lingemann blackness level in field.

Need to prove that the method for training SVM can adopt method of the prior art.

The method that realizes is specifically as follows: for example, the chimney historical data of acquisition has 200 parts, preceding 100 parts of chimney data can be made up recognition network as training sample, as test sample book recognition result is assessed for back 100 parts.Choose the optimum kernel function of identification.

In embodiments of the present invention, the SVM model formation is:

$f\left(x\right)=\mathrm{sgn}(\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{a}_{i}K(X_i,X)+b)---\left(7\right)$

Wherein, sgn (●) is a sign function, $\mathrm{sgn}\left(x\right)=\left\{\begin{array}{c}-1,x<0\\ 0,x=0\\ +1,x>0\end{array}\right.---\left(8\right)$

$K(X_i,X)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}\min (X_i,X)---\left(9\right)$

K (X _i, X) be the common factor kernel function.

In embodiments of the present invention, be 8 grades (0 grades, 1 grade, 2 grades, 3 grades, 3.5 grades, 4 grades, 4.5 grades, 5 grades) with sample classification, certainly in actual conditions, also Lin Geman 0-5 level can be divided into other umber, for example 10 parts or 20 parts.Selected for use to be output as many classification SVM models of 8, many-valued SVM model is realized by making up a plurality of two-value sub-classifiers, needs design 8* (8-1)/2=28 SVM model altogether, if f in the formula (7) _Ij(x)＞0, then declare x and belong to the i class, the i class gets a ticket, otherwise belongs to the j class, and the j class gets a ticket, at last according to voting results, and classification under judging.

In embodiments of the present invention, with 0 grade, 1 grade, 2 grades, 3 grades, 3.5 grades, 4 grades, 4.5 grades, 5 grades are labeled as 0,1,2,3,4,5,6,7 respectively, last according to the label that is divided, and draw corresponding lingemann blackness level.

Optionally, after S15, can also comprise: when blackness value surpasses the blackness threshold value, carry out alarming processing.Concrete alarming processing can for: start alarm module (for example hummer or alarm), the mode remote alarm that can certainly remind by note.

The method that the specific embodiment of the invention provides realizes the intellectuality that blackness detects owing to be that the image of gathering is operated the calculating of finishing blackness value accordingly, has overcome the defective of the subjectivity of artificial detection, improves the advantage that blackness detects.

The present invention also provides a kind of detection system of atmosphere pollution, and this system comprises as shown in Figure 5:

Image acquisition converting unit 51 is used to gather the image of atmosphere pollution, converts the video image that collects to gray level image;

Comparing unit 52, the selected target area that is used for more previously selected reference zone and gray level image determines whether this target area is the plume zone;

Compensating unit 53 is used for after being defined as the plume zone, according to the relation in reference zone and plume zone, the plume zone is carried out the compensation of brightness;

Statistical computation unit 54 is used for statistics of histogram is carried out in the plume zone after the compensation, calculates the center of gravity of grey level histogram;

Computing unit 55 is used for drawing according to the center of gravity calculation of grey level histogram the blackness value in plume zone.

Optionally, comparing unit 52 specifically comprises:

Model building module 521 is used to set up the background model of reference zone, specifically is established as:

$\mathrm{\&mu;}(i,j)=\frac{1}{N}(I_1(i,j)+I_2(i,j)+...+I_N(i,j))$

$\mathrm{\&sigma;}(i,j)=\frac{1}{N}((I_1(i,j)-\mathrm{\&mu;}(i,j))+(I_2(i,j)-\mathrm{\&mu;}(i,j))+...+(I_N(i,j)-\mathrm{\&mu;}(i,j)))$

Wherein, I _N((i j) is the average of reference zone to μ, and (i j) is the reference zone variance to σ for i, j) the reference zone pixel value of expression N frame;

Pixel judge module 522 is used for basis

$D(i,j)=\left\{\begin{array}{c}255,\mathrm{abs}(\mathrm{lr}(i,j)-\mathrm{\&mu;}(i,j))>=N_1\&CenterDot;\mathrm{\&sigma;}(i,j)\\ 0,\mathrm{abs}(\mathrm{lr}(i,j)-\mathrm{\&mu;}(i,j))<N_1\&CenterDot;\mathrm{\&sigma;}(i,j)\end{array}\right.$ Calculate the value after the binaryzation;

Wherein, (i j) is the target area pixel to lr; (i j) is value after the binaryzation to D;

As D (i, j)=255 o'clock, this current pixel point is the smog pixel, otherwise is non-smog pixel;

Statistics judge module 523 is used to add up the smog pixel number in the target area, surpasses threshold value as the smog pixel in the target area, determines that then this target area is the plume zone, otherwise is non-plume zone.

Optionally, compensating unit 53 can comprise:

Mean flow rate computing module 531 is used for basis

$\mathrm{Lr}=\frac{1}{M*N*n}\underset{j=1}{\overset{M}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}(I_1(i,j)+...I_n(i,j))$ Calculate the mean flow rate in the setting-up time internal reference zone;

Wherein, M is the reference zone width, and N is the reference zone height, and n is the reference zone in the statistics n frame; Lr is the mean flow rate of reference zone;

Luminance compensation module 532, be used for according to Ld (i, j)=lr (i, j)-Lr carries out luminance compensation to the plume zone;

Wherein, (i is to the pixel value behind the plume regional compensation j) to Ld.

Optionally, statistical computation unit 54 comprises:

Piecemeal module 541 is used for the plume zone after the compensation is divided into the piece of 3*3;

Statistical module 542 is used for each image subblock is carried out the grey level histogram hist that gray-scale statistical obtains each height piece correspondence ₁[n] ..., hist ₉[n]; Wherein the span of n is 0-255;

Center of gravity calculation module 543 is used for basis

Calculate the center of gravity of the grey level histogram of each sub-piece;

Hist wherein _iThe grey level histogram of i sub-piece of [n] expression; Hist_x _iThe grey level histogram center of gravity of representing i sub-piece.

Optionally, aforementioned calculation unit 55 specifically is used for the center of gravity of each sub-piece of grey level histogram is calculated the blackness level as the good support vector machine SVM of input vector input training in advance.

The system that the specific embodiment of the invention provides realizes the intellectuality that blackness detects owing to be that the image of gathering is operated the calculating of finishing blackness value accordingly, has overcome the defective of the subjectivity of artificial detection, improves the advantage that blackness detects.

It should be noted that said system, each included unit is just divided according to function logic, but is not limited to above-mentioned division, as long as can realize function corresponding; In addition, the concrete title of each functional unit also just for the ease of mutual differentiation, is not limited to protection scope of the present invention.

In addition, one of ordinary skill in the art will appreciate that all or part of step that realizes in the foregoing description method is to instruct relevant hardware to finish by program, corresponding program can be stored in a kind of computer-readable recording medium, the above-mentioned storage medium of mentioning can be a ROM (read-only memory), disk or CD etc.

In sum, technical scheme provided by the invention has the intellectuality that realizes that blackness detects, and has overcome the defective of the subjectivity of artificial detection, improves the advantage that blackness detects.

The above only is preferred embodiment of the present invention, not in order to restriction the present invention, all any modifications of being done within the spirit and principles in the present invention, is equal to and replaces and improvement etc., all should be included within protection scope of the present invention.

Claims (10)

1. the detection method of an atmosphere pollution is characterized in that, described method comprises the steps:

Gather the image of atmosphere pollution, convert the video image that collects to gray level image;

Selected target area in more previously selected reference zone and the gray level image determines whether this target area is the plume zone;

As after being defined as the plume zone,, the plume zone is carried out the compensation of brightness according to the relation in reference zone and plume zone;

Statistics of histogram is carried out in plume zone after the compensation, calculate the center of gravity of grey level histogram;

Draw the blackness value in plume zone according to the center of gravity calculation of grey level histogram.

2. method according to claim 1 is characterized in that, the selected target area in described more previously selected reference zone and the gray level image determines that whether this target area is that the step in plume zone specifically comprises:

Set up the background model of reference zone, the concrete method of setting up is:

$\mathrm{\&mu;}(i,j)=\frac{1}{N}(I_1(i,j)+I_2(i,j)+...+I_N(i,j))$

$\mathrm{\&sigma;}(i,j)=\frac{1}{N}((I_1(i,j)-\mathrm{\&mu;}(i,j))+(I_2(i,j)-\mathrm{\&mu;}(i,j))+...+(I_N(i,j)-\mathrm{\&mu;}(i,j)))$

Wherein, I _N((i j) is the average of reference zone to μ, and (i j) is the reference zone variance to σ for i, j) the reference zone pixel value of expression N frame; (i, j) the capable j of the i of presentation video is listed as respectively;

According to

Calculate the value after the binaryzation;

Wherein, (i j) is the target area pixel to lr; (i j) is value after the binaryzation to D;

When D (i, j)=255 o'clock, this current pixel point is the smog pixel, otherwise is non-smog pixel;

Smog pixel number in the statistics target area surpasses threshold value as the smog pixel in the target area, determines that then this target area is the plume zone, otherwise is non-plume zone.

3. method according to claim 1 is characterized in that, the relation in described reference zone and plume zone, and the step of the plume zone being carried out the compensation of brightness specifically comprises:

According to

Calculate the mean flow rate in the setting-up time internal reference zone;

Wherein, M is the reference zone width, and N is the reference zone height, and n is the reference zone in the statistics n frame; Lr is the mean flow rate of reference zone;

According to Ld (i, j)=lr (i, j)-Lr carries out luminance compensation to the plume zone;

Wherein, (i is to the pixel value behind the plume regional compensation j) to Ld.

4. method according to claim 1 is characterized in that, described statistics of histogram is carried out in plume zone after the compensation, and the step of calculating the center of gravity of grey level histogram comprises:

The piece that plume zone after the compensation is divided into 3*3;

Each image subblock is carried out the grey level histogram hist that gray-scale statistical obtains each height piece correspondence ₁[n] ..., hist ₉[n]; Wherein the span of n is 0-255;

According to

Calculate the center of gravity of the grey level histogram of each sub-piece;

Hist wherein _iThe grey level histogram of i sub-piece of [n] expression; Hist_x _iThe grey level histogram center of gravity of representing i sub-piece.

5. method according to claim 1 is characterized in that, the step that described center of gravity calculation according to grey level histogram draws the blackness value in plume zone specifically comprises:

The center of gravity of each sub-piece of grey level histogram is calculated the blackness level as the good support vector machine SVM of input vector input training in advance.

6. the detection system of an atmosphere pollution is characterized in that, described system comprises:

The image acquisition converting unit is used to gather the image of atmosphere pollution, converts the video image that collects to gray level image;

Comparing unit, the selected target area that is used for more previously selected reference zone and gray level image determines whether this target area is the plume zone;

Compensating unit is used for after being defined as the plume zone, according to the relation in reference zone and plume zone, the plume zone is carried out the compensation of brightness;

The statistical computation unit is used for statistics of histogram is carried out in the plume zone after the compensation, calculates the center of gravity of grey level histogram;

Computing unit is used for drawing according to the center of gravity calculation of grey level histogram the blackness value in plume zone.

7. system according to claim 6 is characterized in that, described comparing unit specifically comprises:

Model building module is used to set up the background model of reference zone, specifically is established as:

$\mathrm{\&mu;}(i,j)=\frac{1}{N}(I_1(i,j)+I_2(i,j)+...+I_N(i,j))$

$\mathrm{\&sigma;}(i,j)=\frac{1}{N}((I_1(i,j)-\mathrm{\&mu;}(i,j))+(I_2(i,j)-\mathrm{\&mu;}(i,j))+...+(I_N(i,j)-\mathrm{\&mu;}(i,j)))$

Wherein, I _N((i j) is the average of reference zone to μ, and (i j) is the reference zone variance to σ for i, j) the reference zone pixel value of expression N frame; (i, j) the capable j of the i of presentation video is listed as respectively;

The pixel judge module is used for basis

Calculate the value after the binaryzation;

Wherein, (i j) is the target area pixel to lr; (i j) is value after the binaryzation to D;

When D (i, j)=255 o'clock, this current pixel point is the smog pixel, otherwise is non-smog pixel;

The statistics judge module is used to add up the smog pixel number in the target area, surpasses threshold value as the smog pixel in the target area, determines that then this target area is the plume zone, otherwise is non-plume zone.

8. system according to claim 6 is characterized in that, described compensating unit comprises:

The mean flow rate computing module is used for basis

Calculate the mean flow rate in the setting-up time internal reference zone;

Wherein, M is the reference zone width, and N is the reference zone height, and n is the reference zone in the statistics n frame; Lr is the mean flow rate of reference zone;

The luminance compensation module, be used for according to Ld (i, j)=lr (i, j)-Lr carries out luminance compensation to the plume zone;

Wherein, (i is to the pixel value behind the plume regional compensation j) to Ld.

9. system according to claim 6 is characterized in that, described statistical computation unit comprises:

The piecemeal module is used for the plume zone after the compensation is divided into the piece of 3*3;

Statistical module is used for each image subblock is carried out the grey level histogram hist that gray-scale statistical obtains each height piece correspondence ₁[n] ..., hist ₉[n]; Wherein the span of n is 0-255;

Center of gravity calculation module is used for basis

Calculate the center of gravity of the grey level histogram of each sub-piece;

Hist wherein _iThe grey level histogram of i sub-piece of [n] expression; Hist_x _iThe grey level histogram center of gravity of representing i sub-piece.

10. system according to claim 6 is characterized in that, described computing unit specifically is used for the center of gravity of each sub-piece of grey level histogram is calculated the blackness level as the good support vector machine SVM of input vector input training in advance.

Images