CN104020088B - A kind of method obtaining particulate in air concentration based on image procossing - Google Patents

Abstract

The invention discloses a kind of method obtaining particulate in air concentration based on image procossing, the image gathered processed, comprises the following steps: 1) determine the pixel value of atmosphere light pixel the brightest in image and relative dark interest pixel；2) according to pixel value and the pixel value of the interest pixel determined of the atmosphere light pixel obtained, the light intensity value of respective pixel point is calculated respectively；3) according to step 2) light intensity value of respective pixel point that obtains calculates light intensity attenuation coefficient；4) based on Mie theory, the extinction coefficient relational expression function about the mean diameter of particulate matter is determined；5) occurrence of the mean diameter of particulate matter is determined；6) particle concentration is calculated according to the mean diameter of light intensity attenuation coefficient, the relational expression function of extinction coefficient and particulate matter.The method obtaining particulate in air concentration based on image procossing of the present invention, it is possible to realize the estimation to the particle concentration in the air in wider range at lower cost.

Description

A kind of method obtaining particulate in air concentration based on image procossing

[technical field]

The present invention relates to computer vision field, particularly relate to a kind of method obtaining particulate in air concentration based on image procossing.

[background technology]

Along with industry and the development of transportation, substantial amounts of harmful substance is discharged in air, changes the normal composition of air, makes air quality degenerate.When people live among the air being contaminated, not only normal life and trip also can be by strong influences, and the health of human body can be encroached on greatly.According to it has been reported that haze weather have resulted in China many suffer severe contamination, on people's productive life with healthy bring extremely disadvantageous impact.

In order to reduce the infringement of haze, and it is provided with the information helping administer, it is necessary to set up comprehensive air quality monitoring network network.At present, relatively more authoritative have national environmental protection department by setting up substantial amounts of observation site and the air quality information of real-time broadcasting.Single monitoring network needs to be equipped with the concussion professional equipment such as balance, Beta alpha cellulose a gage and measures PM value in real time, and per hour to Web realease measurement data.But, current monitoring net there is also a lot of limitation.First, such as, shaking the professional equipment such as balance, Beta alpha cellulose a gage, it costs dearly, operation complexity, is unfavorable for promoting.Secondly, in a city, the quantity of monitoring network is the most limited, and the effective range that single site is measured is limited, to such an extent as to bed rearrangement city monitoring coverage and limited.A lot of regions are away from monitoring network, and air quality situation is the most complicated and changeable so that the data of site do not possess referential to these regional air quality-monitorings.

[summary of the invention]

The technical problem to be solved is: make up above-mentioned the deficiencies in the prior art, a kind of method obtaining particulate in air concentration based on image procossing is proposed, it is possible to realize the estimation to the particle concentration in the air in wider range at lower cost and obtain.

The technical problem of the present invention is solved by following technical scheme:

A kind of method obtaining particulate in air concentration based on image procossing, the image that collection includes outdoor scenes processes, and comprises the following steps: 1) determine the pixel value of atmosphere light pixel the brightest in image and relative dark interest pixel；2) according to pixel value and the pixel value of the interest pixel determined of the atmosphere light pixel obtained, the light intensity value of respective pixel point is calculated respectively；3) according to step 2) light intensity value of respective pixel point that obtains calculates light intensity attenuation coefficient；4) based on Mie theory, the extinction coefficient relational expression function about the mean diameter of particulate matter is determined；5) occurrence of the mean diameter of particulate in air is determined according to the relational expression function of light intensity attenuation coefficient and extinction coefficient；6) particle concentration is calculated according to the mean diameter of light intensity attenuation coefficient, the relational expression function of extinction coefficient and particulate matter.

The present invention is compared with the prior art and provides the benefit that:

The method obtaining particulate in air concentration based on image procossing of the present invention, based on image processing techniques, extracts the light intensity value of respective pixel point in image, thus obtains light intensity attenuation coefficient, in conjunction with extinction coefficient, and then be finally calculated the mean diameter of particulate matter.In the method for the acquisition air quality parameters of the present invention, it is only necessary to a device gathering image, it is not necessary to expensive professional equipment, the cost that method realizes is relatively low, it is easy to promote.Simultaneously, the sub-picture only needing shooting, collecting region to be monitored can obtain air quality parameters information, i.e. for geographical position inconvenience, the region away from observation site is also easy to use, i.e. the method for the present invention can realize the monitoring to the air ambient in wider range.

[accompanying drawing explanation]

Fig. 1 is the flow chart of the method obtaining particulate in air concentration based on image procossing of the specific embodiment of the invention；

Fig. 2 is determination air pixel and the flow chart of interest pixel step of the specific embodiment of the invention；

Fig. 3 is the schematic diagram of outdoor scene carrying out in the specific embodiment of the invention choosing during experiment test；

The curve comparison figure of the observation at estimated value when Fig. 4 is to carry out experiment test in the specific embodiment of the invention and actual observation station；

The relevant figure of estimated value when Fig. 5 is to carry out experiment test in the specific embodiment of the invention and the observation at actual observation station.

[detailed description of the invention]

Below in conjunction with detailed description of the invention and compare accompanying drawing the present invention is described in further details.

The present invention, by image procossing and the technology of identification, processes the image of the image capturing device such as video camera or photographing unit shooting, and by the particulate in air concentration information of image, i.e. PM value extracts.Associating in current large-and-medium size cities, suffer from a large amount of fixing CCTV cameras, be covered with in city region greatly, therefore the method for the present invention is suitable for large-area popularization and is suitable for.For CCTV camera, as long as photographing a range of outdoor scenes image, PM value in scene can be obtained in the range of this.And in a city, the quantity of CCTV camera is far longer than the monitoring site of specialty at present, collects these data, a finest PM Distribution value figure in city can be obtained.Thering is provided more detailed air pollution distribution situation for i.e. relevant unit of governments at all levels, environmental administration, formulation and execution for every decision-making provide reference more fully and effectively.

As it is shown in figure 1, be the flow chart of the method obtaining particulate in air concentration in this detailed description of the invention based on image procossing.First, collecting a width and include the image of outdoor scenes, this image is directly to obtain under normal photographing, it is not necessary to process through other technologies means (such as exposure correction, AWB etc.).Then processing this image, processing procedure includes:

P1) pixel value of atmosphere light pixel the brightest in image and relative dark interest pixel are determined.

In this detailed description of the invention, use steps flow chart as shown in Figure 2 to determine above-mentioned two parts content, thus degree of accuracy is higher, contributes to guaranteeing light intensity attenuation coefficient, the mean diameter of particulate matter and the degree of accuracy of particle concentration that subsequent calculations obtains.Other also has some methods quickly determining two parts content, but counting accuracy can be a bit weaker, numerous to list herein.In this detailed description of the invention, as in figure 2 it is shown, include:

P11) brightness value of each pixel in image is calculated.

In this step, the brightness value calculating each pixel has multiple method.A kind of computational methods faster particularized below: be that (x, the brightness value of pixel y) is I for position^light(x, y),

I^light(x, y)=0.299 × I^R(x,y)+0.587×I^G(x,y)+0.114×I^B(x,y)

Wherein, I^R(x, y), I^G(x, y), I^B(x y) is respectively pixel (x, R passage pixel value y), G passage pixel value, channel B pixel value.

The above-mentioned computational methods enumerated are only the one in multiple computational methods, and remaining computational methods may have more complicated mode, and its operation result is more accurate, but there do not have said method to calculate like that to be easy to be quick.Which kind of calculation specifically chosen, user can consider selected according to the requirement of the degree of accuracy to the efficiency of processing procedure and result.

P12) locally darkening processes, and obtains changing image.Particularly as follows: the regional area determined around each pixel, obtain the minimum luminance value in the brightness value of pixel in described regional area, described minimum luminance value is defined as the conversion brightness value of pixel, it is thus achieved that changing image.

In this step, the regional area determined can be arbitrary shape, it is therefore preferable to rectangular window, and the size of rectangular window is M × N, and wherein the concrete value of M and N is set by the user.

P13) pixel value of atmosphere light pixel is calculated, particularly as follows: in described step 12) in the changing image that obtains, finding brightness value and account for the high luminance pixel point of front p%, p is value between 5～15, using the pixel value average of described high luminance pixel point as the pixel value of atmosphere light pixel.

This step is to estimate the pixel value of atmosphere light pixel, previous step, in the conversion changing image that constitutes of brightness value determined, search and account for the pixel of p% high luminance values before all pixels.For the value of p, value between 5～15.After determining this partial pixel point, take the average of pixel value in the image that this partial pixel point is original as the pixel value of atmosphere light pixel, be designated as, wherein { R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B to c ∈ respectively.Namely, the average of R passage pixel value in the pixel original image that pixel value is the above-mentioned part determined of atmosphere light pixel R passage, the average of R passage pixel value in the pixel original image that pixel value is the above-mentioned part determined of atmosphere light pixel R passage, the average of R passage pixel value in the pixel original image that pixel value is the above-mentioned part determined of atmosphere light pixel R passage.

P14) interest pixel is determined, particularly as follows: the pixel meeting following condition to be defined as the most dark interest pixel: (I^light(x,y)-I^minlight(x,y))²＜ ε², wherein, I^light(x y) is the brightness value of pixel, I^minlight(x, y) is the conversion brightness value of pixel, and ε is the threshold value that user sets.

This step is to choose the most dark pixel required for subsequent calculations light intensity attenuation coefficient.To pixels all in image, choose point of interest by following criterion: (I^light(x,y)-I^minlight(x,y))²＜ ε²If above expression formula can be met, then this point is as interest pixel.Statistics interest pixel quantity n=1,2,3 ..., N.Record interest pixel position is (x_n, y_n), and record the interest pixel sum N obtained.

As above, the pixel value of atmosphere light pixel, and interest pixel are i.e. determined by above-mentioned steps.Above-mentioned determine process, it is that the content for full image carries out overall merit, be equivalent to the observation having the order of magnitude suitable with camera pixel number, thus pixel value and the interest pixel of atmosphere light pixel is obtained according to the pixel assessment of entire image, the result that assessment determines is more accurate, can effectively reduce because single-point assesses introduced error, and then contribute to guaranteeing the degree of accuracy of the result of calculation of follow-up light intensity dampening information and particle concentration.

After determining above-mentioned two parts content, enter step P2).

P2) according to pixel value and the pixel value of the interest pixel determined of the atmosphere light pixel obtained, the light intensity value of respective pixel point is calculated respectively.

In this step, pixel value is converted to light intensity value information, for the calculating of follow-up light intensity attenuation quotient.The calculated for pixel values of the respective channel according to atmosphere light pixel obtains the light intensity value of the respective channel of atmosphere light pixel, obtains the light intensity value of the respective channel of interest pixel according to the calculated for pixel values of the respective channel of interest pixel.When specifically calculating, can be calculated according to equation below: R^c=f^c(I^c), wherein, { R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B, R to c ∈ respectively^cRepresent the light intensity value of the c passage of pixel to be calculated, I^cRepresent the pixel value of the c passage of pixel to be calculated, f^cRepresent the mapping relations between c passage light intensity value and pixel value, the performance parameter of the filming apparatus gathering image determine.Preferably the corresponding mapping relations between light intensity to pixel value are linear.But it practice, affected by the performance parameter of the filming apparatus gathering image, not being preferable linear relationship, different cameras often has different mapping response curves, can be determined the mapping relations of this camera by experimental calibration binding ability parameter.

P3) according to step P2) light intensity value of respective pixel point that obtains calculates light intensity attenuation coefficient.

In this step, calculate light intensity attenuation coefficient by the light intensity value of atmosphere light pixel and interest pixel in image.In this detailed description of the invention, expression formula as follows is used to calculate light intensity attenuation coefficient:

${\mathrm{\τ}}^c=\underset{n=1}{\overset{N}{\mathrm{\Σ}}}\frac{\ln \frac{R_{\∞}^c-R^c(x_n,y_n)}{R_{\∞}^c}}{d(x_n,y_n)}$

Wherein, c ∈ R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B respectively,Representing the light intensity value of the c passage of atmosphere light pixel, N represents total number of point of interest, R^c(x_n,y_n) represent interest pixel (x_n, y_n) the light intensity value of c passage, d (x_n, y_n) represent interest pixel (x_n, y_n) distance between corresponding point in kind and the filming apparatus gathering image；τ^cRepresent the light intensity attenuation coefficient of c passage.By calculating, the final light intensity attenuation coefficient τ obtaining tri-Color Channels of RGB^R、τ^G、τ^B。

P4) based on Mie theory, the extinction coefficient relational expression function about the mean diameter of particulate matter is determined.

Extinction coefficient can be calculated by Mie theory obtain with wavelength X c, particulate matter mean diameter S, particulate matter refractive index m for input.Based on Mie theory, there is more computing formula can calculate extinction coefficient.The formula of the most fast and convenient a kind of determination extinction coefficient particularized below:

$K_c\left(S\right)=(1-e^{-\frac{\mathrm{\α}}{A}})[2+B({\left(\frac{\mathrm{\α}}{C}\right)}^2\exp (-{\left(\frac{\mathrm{\α}}{C}\right)}^2)+\frac{\frac{\mathrm{\α}}{4C}}{{\left(\frac{\mathrm{\α}}{C}\right)}^{1.5}+1.5})-\frac{\cos \left((E-\mathrm{Fexp}(-\frac{\mathrm{\α}}{2C}))\mathrm{\α}\right)}{{\left(\mathrm{D\α}\right)}^{1.3}+0.5}]$

Wherein, c ∈ R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B respectively, $A=\frac{0.5}{n-1+{3\mathrm{\η}}^2},$ $B=1.18n^2{(n-1)}^{0.1}\exp \{-[\frac{\mathrm{\ηn}}{n-1}-(n-1)D\left]\right\},$ $C=\frac{2(1-\mathrm{\η})}{n-1},$ $D=0.8231(\frac{5n-3}{3n-1}+10\frac{\sqrt{\mathrm{\η}}}{n}),$ E=0.7 (n-1) (1.7+n^0.13), F=0.7 (n-1) n^0.13, $\mathrm{\α}=\frac{\mathrm{\πS}}{{\mathrm{\λ}}_c},$ Wherein, n and η represents real part and the imaginary part of particulate matter refractive index m respectively, and n is value between 1.3～1.9, and η is value between 0.01～0.1, and S represents that the mean diameter of particulate matter, λ c represent the optical wavelength of c passage light.

In above-mentioned coefficient, optical wavelength λ c determines that known.Particulate matter refractive index m generally plural number, is designated as n+i η, n and η and is set by the user, and can suitably adjust setting with the shooting on-site difference of image.After setting, above-mentioned each parameter is brought in formula, i.e. obtain extinction coefficient K under c passage_cRelational expression function K about mean diameter S of particulate matter_c(S), the relational expression function K under concrete i.e. tri-Color Channels of RGB_R(S)、K_G(S)、K_B(S)。

P5) occurrence of the mean diameter of particulate in air is determined according to the relational expression function of light intensity attenuation coefficient and extinction coefficient.

This step is according to the mean diameter of the relational expression Function Estimation particulate in air impurity of light intensity attenuation coefficient and extinction coefficient.In this detailed description of the invention, method of least square is used to determine the occurrence of mean diameter.Specifically, by the light intensity attenuation coefficient τ of each passage^R、τ^G、τ^BAnd the relational expression function K of extinction coefficient_R(S)、K_G(S)、K_B(S) following expression is brought into: ${[\frac{{\mathrm{\τ}}^R}{{\mathrm{\τ}}^G}-\frac{K_R\left(S\right)}{K_G\left(S\right)}]}^2+{[\frac{{\mathrm{\τ}}^R}{{\mathrm{\τ}}^B}-\frac{K_R\left(S\right)}{K_B\left(S\right)}]}^2+[\frac{{\mathrm{\τ}}^G}{{\mathrm{\τ}}^B}-\frac{K_G\left(S\right)}{K_B\left(S\right)}{]}^2,$ Calculate the value of the S of the value minimum making described relational expression, be i.e. calculated the occurrence of the mean diameter of particulate matter.

P6) particle concentration is calculated according to the mean diameter of light intensity attenuation coefficient, the relational expression function of extinction coefficient and particulate matter.

Light intensity attenuation coefficient is according to step P3) obtain, the mean diameter of particulate matter is according to step P5) determine obtain, by step P5) determine that the occurrence of the mean diameter of the particulate matter obtained substitutes into step P4) the relational expression function that determines is calculated the occurrence of extinction coefficient under each passage.Known light intensity attenuation coefficient, extinction coefficient, mean diameter, according to itself and the relational expression of particle concentration, particle concentration can be calculated.

Specifically, can be according to following expression calculating particle concentration P:Wherein { R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B, τ to c ∈ respectively^cRepresent the light intensity attenuation coefficient of c passage, K_c(S) extinction coefficient K under c passage is represented_cAbout the relational expression function of mean diameter S of particulate matter, S represents the mean diameter of particulate matter.Substitute into light intensity attenuation coefficient, extinction coefficient about the function of particulate matter mean diameter and the occurrence of the mean diameter of particulate matter, i.e. can determine that and obtain particle concentration.

After obtaining particle concentration, can also be according to the mean diameter of particulate matter and particle concentration, " ambient air quality " (GB3095-2012) that inquiry country promulgates, judges to air quality, obtains the air quality information of environment in the image gathered.

To sum up, i.e. according to image processing techniques, obtain the concentration of the particulate in air of the outdoor scene of shooting.In the method for the acquisition air quality parameters of this detailed description of the invention, use the brand-new design being obtained air quality information by image procossing, it is different from traditional air quality measuring method, as: concussion balance weight method, laser scattering method etc., this method utilizes the image that general camera shoots, based on image processing techniques, extract the anti-particle concentration pushed away in scene of feature characterizing haze degree in image.Only needing a device gathering image in method, it is not necessary to other sensing devices, without expensive professional equipment, the cost that method realizes is relatively low, it is easy to promote.Application this method, a common camera or general camera, a real-time Detection of Air Quality point can be become, it is not necessary to the time as longer when single measurement in existing air quality monitoring stations gathers air.Meanwhile, it is also convenient for being adapted to diverse geographic location, the atmospheric environment of different mechanism, the monitoring to the air ambient in wide scope can be realized.

As follows, apply the method for this detailed description of the invention to carry out many experiments test.During test, choose outdoor multiple images of scene capture 70 in middle section's resource mansion of Haidian District Beijing shown in Fig. 3 (typically choosing shooting on daytime, shoot 4 images respectively in a day).Processing multiple image, the parameter of setting is as follows: setting the window size window as 25 × 25 when local darkening processes, parameter p takes 10, determines ε during interest pixel²=10, the light intensity that shooting camera is determined with the function of pixel value is: R^R=0.321 × I^R-15.52, R^G=0.276 × I^G-12.99, R^B=0.339 × I^B-15.04, extract the range information between point in kind and the shooting camera of the information acquisition point of interest reaction in the depth map of image, take particulate matter refractive index m=1.5+0.01i, i.e. n=1.5, η=0.01, λ_R=700nm, λ_G=546.1nm, λ_B=435.8nm.Under the above parameters, 70 multiple images to this scene capture carry out the estimation of particle concentration (with PM2.5 as parameter index), the PM2.5 data observation value that the estimated result obtained and environmental protection part are observed in the observation station of Web realease contrasts, and obtains curve chart as shown in Figure 4.Relation between the estimated result obtained and the observation of reality is depicted as picture and obtains dot chart as shown in Figure 5.From Fig. 4 and Fig. 5 it is recognized that while the estimated value of the method for estimation of this detailed description of the invention cannot guarantee that 100% ground is accurate, but with actual observed value very close to, it was predicted that value and observation correlation coefficient reach 0.8686, it was predicted that mean square error be ± 35.6830 μ g/m³.The method applying this detailed description of the invention, can carry out observability estimate fast and effectively to the particle concentration in environment.

Above content is to combine concrete preferred implementation further description made for the present invention, it is impossible to assert the present invention be embodied as be confined to these explanations.For general technical staff of the technical field of the invention, make some replacements or obvious modification without departing from the inventive concept of the premise, and performance or purposes are identical, all should be considered as belonging to protection scope of the present invention.

Claims (9)

1. the method obtaining particulate in air concentration based on image procossing, it is characterised in that: collection is comprised The image having outdoor scenes processes, and comprises the following steps:

1) pixel value of atmosphere light pixel the brightest in image and relative dark interest pixel are determined；

2) according to pixel value and the pixel value of the interest pixel determined of the atmosphere light pixel obtained, phase is calculated respectively Answer the light intensity value of pixel；

3) according to step 2) light intensity value of respective pixel point that obtains calculates light intensity attenuation coefficient；

4) based on Mie theory, the extinction coefficient relational expression function about the mean diameter of particulate matter is determined；

5) tool of the mean diameter of particulate in air is determined according to the relational expression function of light intensity attenuation coefficient and extinction coefficient Body value；

6) particulate matter is calculated according to the mean diameter of light intensity attenuation coefficient, the relational expression function of extinction coefficient and particulate matter Concentration.

The method obtaining particulate in air concentration based on image procossing the most according to claim 1, its feature exists In described step 1) in specifically include following steps: 11) calculate the brightness value of each pixel in image；12) Determining the regional area around each pixel, the minimum obtained in described regional area in the brightness value of pixel is bright Angle value, is defined as the conversion brightness value of pixel, it is thus achieved that changing image by described minimum luminance value；13) in described step Rapid 12) in the changing image obtained, finding brightness value and account for the high luminance pixel point of front p%, p is value between 5～15, Using the pixel value average of described high luminance pixel point as the pixel value of atmosphere light pixel；14) following condition will be met Pixel be defined as the most dark interest pixel: (I^light(x,y)-I^minlight(x,y))²<ε², wherein, I^light(x y) is the brightness value of pixel, I^minlight(x, y) is the conversion brightness value of pixel, and ε is that user sets Threshold value.

The method obtaining particulate in air concentration based on image procossing the most according to claim 2, its feature exists In described step 12) in sized by the regional area that determines for the rectangular area of M × N, the concrete value of M and N It is set by the user.

The method obtaining particulate in air concentration based on image procossing the most according to claim 1, its feature exists In described step 2) according to equation below by the light intensity value of the calculated for pixel values pixel of pixel: R^c=f^c(I^c), Wherein, c ∈ R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B respectively, I^cRepresent the pixel value of the c passage of pixel to be calculated, f^cRepresent the mapping relations between c passage light intensity value and pixel value, Determined by the performance parameter of the filming apparatus gathering image, R^cRepresent the light intensity value of the c passage of pixel to be calculated.

The method obtaining particulate in air concentration based on image procossing the most according to claim 1, its feature exists In described step 3) in calculate light intensity attenuation coefficient according to following expression, Wherein, c ∈ R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B respectively, Representing the light intensity value of the c passage of atmosphere light pixel, N represents total number of point of interest, R^c(x_n,y_n) represent emerging Interest point (x_n, y_n) the light intensity value of c passage, d (x_n, y_n) represent point of interest (x_n, y_n) corresponding point in kind with Gather the distance between the filming apparatus of image；τ^cRepresent the light intensity attenuation coefficient of c passage.

The method obtaining particulate in air concentration based on image procossing the most according to claim 1, its feature exists In described step 4) in the relational expression function of extinction coefficient that determines be:

$K_c\left(S\right)=\left(1-e^{-\frac{\mathrm{\&alpha;}}{A}}\right)\&lsqb;2+B\left({\left(\frac{\mathrm{\&alpha;}}{C}\right)}^2\exp \left(-{\left(\frac{\mathrm{\&alpha;}}{C}\right)}^2\right)+\frac{\frac{\mathrm{\&alpha;}}{4C}}{{\left(\frac{\mathrm{\&alpha;}}{C}\right)}^{1.5}+1.5}\right)-\frac{\cos \left(\left(E-F\exp \left(-\frac{\mathrm{\&alpha;}}{2C}\right)\right)\mathrm{\&alpha;}\right)}{{\left(D\mathrm{\&alpha;}\right)}^{1.3}+0.5}\&rsqb;,$

Wherein, { R, G, B}, R, G, B represent that the R passage of rgb color space, G passage, B are logical to c ∈ respectively Road, $A=\frac{0.5}{n-1+3{\mathrm{\&eta;}}^2},$ $B=1.18n^2{(n-1)}^{0.1}\exp \{-\&lsqb;\frac{\mathrm{\&eta;}n}{n-1}-(n-1)D\&rsqb;\},$ $C=\frac{2(1-\mathrm{\&eta;})}{n-1},$ E=0.7 (n-1) (1.7+n^0.13), F=0.7 (n-1) n^0.13,Its In, n and η represents real part and the imaginary part of particulate matter refractive index respectively, and n is value between 1.3～1.9, and η is 0.01～0.1 Between value, S represents that the mean diameter of particulate matter, λ c represent the optical wavelength of c passage light, K_c(S) represent that c leads to Extinction coefficient K under road_cRelational expression function about mean diameter S of particulate matter.

The method obtaining particulate in air concentration based on image procossing the most according to claim 1, its feature exists In: described step 5) in, by the light intensity attenuation coefficient τ of each passage^R、τ^G、τ^BAnd the relational expression function of extinction coefficient K_R(S)、K_G(S)、K_B(S) following expression is substituted into:Calculate the S's of the value minimum making described relational expression Value, is defined as the occurrence of the mean diameter of particulate matter.

The method obtaining particulate in air concentration based on image procossing the most according to claim 1, its feature exists In: described step 6) in, according to following expression calculating particle concentration P:Wherein { R, G, B}, R, G, B represent the R passage of rgb color space, G passage, channel B, τ to c ∈ respectively^cRepresent c The light intensity attenuation coefficient of passage, K_c(S) extinction coefficient K under c passage is represented_cMean diameter S about particulate matter Relational expression function, S represents the mean diameter of particulate matter.

The method obtaining particulate in air concentration based on image procossing the most according to claim 1, its feature exists In described step 6) after also include step 7), according to mean diameter and the particle concentration inquiry air of particulate matter Quality information, obtains the air quality information of the image Scene gathered.