A haze identification method combining satellite remote sensing multispectral and geographic information

Abstract

The present invention discloses a haze identification method combining satellite remote sensing multispectral and geographic information, including obtaining the performance reflectance data and brightness temperature data of at least two preset spectral channels of remote sensing satellite images, sequentially judging the first to fourth preset conditions for each pixel in the remote sensing satellite image, and respectively excluding ice and snow pixels, inland water pixels, cloud pixels and clear sky pixels when the judgments are yes, wherein the clear sky pixels are identified according to the fourth preset condition and the clear sky identification conditions under different altitudes, and the pixels in the remaining area after exclusion are haze pixels. The present invention has the advantages of fast and stable cloud and haze detection, improving the accuracy of cloud and haze identification, etc., has stronger universality, and improves user experience.

Description

Technical Field

[0001] The present invention relates to the field of satellite remote sensing detection technology, and in particular to a haze recognition method combining satellite remote sensing multi-spectral and geographic information. Background Art

[0002] According to the International Meteorological Organization, haze is defined as days when the average daily visibility is less than 10 km, the average daily relative humidity is less than 80%, and other atmospheric haze that can cause low visibility is excluded, such as precipitation. Haze particles are mainly composed of extremely small particles suspended in the air, which are relatively dry compared to fog and relatively wet compared to dust, such as dust, sulfate, nitrate, etc. There is no obvious boundary between haze and clear sky areas, and it looks yellow or orange-gray. The thick haze layer can usually extend to hundreds of meters and last for many days, causing many inconveniences to those engaged in air quality monitoring, public transportation, and physical health. In recent years, the phenomenon of haze weather has increased rapidly. The mutual superposition of haze and cloud layers has brought great challenges to the satellite quantitative remote sensing inversion of cloud characteristics, aerosol characteristics, and the study of the interaction between the two.

[0003] The traditional haze observation method mainly relies on the ground station to observe continuously in a short time through a small range of points. It has obvious advantages in the analysis of the causes of special pollution processes and the analysis of pollutant change trends. However, when it is necessary to identify clouds, clear sky and haze on a large scale, this method loses its advantages and needs to use satellite remote sensing to carry out related work. The traditional method of identifying haze pixels based on satellite remote sensing observation data mainly takes into account the differences in the spectral characteristics of clouds and haze. The traditional threshold method is the most widely used and easy to implement, but it is easy to misclassify clouds and haze. Moreover, this type of method is very sensitive to the threshold of the identification condition, and there is a certain uncertainty in the selection of the threshold. At the same time, the boundary between clear sky and haze area is fuzzy, and only using the threshold method often causes large misjudgment, which introduces more uncertainty to the inversion of aerosol parameters in haze area. In addition, the detection accuracy of clouds and haze in the current heavy pollution weather is poor. Severe haze areas are easily misjudged as clouds, and light haze areas are easily misjudged as clear sky. Its application in real-time business operation is prone to poor universality. Summary of the invention

[0004] The present invention provides a haze recognition method combining satellite remote sensing multispectral and geographic information, further analyzes the spectral characteristics of haze, cloud and clear sky pixels, combines the spectral change rate and altitude information, and uses a tree analysis method to achieve the distinction between ice and snow pixels, inland water pixels, cloud pixels, clear sky pixels and haze pixels, thereby solving the problem of poor recognition accuracy of existing cloud and haze recognition algorithms in heavy pollution weather.

[0005] The present invention comprises the following steps:

[0006] Step 1: Obtain reflectance data ρ of at least two preset spectral channels of remote sensing satellite images λ and brightness temperature data BT, and preprocessing the data to obtain preprocessed data, wherein the at least two preset spectral channels are respectively a visible light channel and an infrared channel;

[0007] Step 2: for each pixel in the remote sensing satellite image, the following judgment is made: whether the preprocessed data meets a first preset condition, and if so, the pixel is marked as an ice and snow pixel;

[0008] Step 3: If the result of step 2 is no, then determine whether the preprocessed data meets the second preset condition; if the result is yes, then mark the pixel as an inland water pixel;

[0009] Step 4: If the result of step 3 is no, then determine whether the preprocessed data meets a third preset condition; if the result of step 3 is yes, then mark the pixel as a cloud pixel;

[0010] Step 5. If the judgment of step 4 is no, then determine whether the preprocessed data meets the fourth preset condition. If it is yes, mark the pixel as a clear sky pixel. If it is no, mark the pixel as a haze pixel, wherein the fourth preset condition includes an altitude geographic information preset condition and a spectral change rate preset condition.

[0011] Further, the preset spectral channels include: a blue light band channel, a red light band channel, a first near infrared channel, a second near infrared channel, a third near infrared channel, a fourth near infrared channel and a long-wave infrared channel;

[0012] The spectral band center wavelength λ of the preset spectral channel is respectively expressed as: blue light band channel λ blue , red light band channel λ red , the first near-infrared channel λ nir1 , the second near-infrared channel λ nir2 , the third near infrared channel λ nir3 , the fourth near-infrared channel λ nir4 and the long-wave infrared channel λ infra , where 0.8μm≤λ nir1 ≤0.9μm, 1.3μm≤λ nir2 ≤1.4μm, 1.6μm≤λ nir3 ≤1.7μm, 2.1μm<λ nir4 ≤2.3μm, 7μm≤λ infra ≤14μm.

[0013] Furthermore, the central wavelength λ of the first near-infrared channel is nir1is 0.865 μm, the central wavelength λ of the second near-infrared channel nir2 is 1.38 μm, the central wavelength λ of the third near-infrared channel nir3 is 1.64 μm, the central wavelength λ of the fourth near-infrared channel nir4 The central wavelength of the long-wave infrared channel is 2.13 μm and λ infra It is 11μm.

[0014] Furthermore, the first preset condition is: And BT infra <t2, where ρ nir1 is the apparent reflectance of the first near-infrared channel, ρ nir3 is the apparent reflectance of the third near-infrared channel, t1 is the first preset threshold, BT infra is the brightness temperature of the long-wave infrared channel, and t2 is the second preset threshold.

[0015] Furthermore, the second preset condition is: And nir4 <t4, where ρ red are the apparent reflectances of the red light band channel respectively, and t3 is the third preset threshold; nir4 is the apparent reflectivity of the fourth near-infrared channel, and t4 is the fourth preset threshold.

[0016] Furthermore, the third preset condition is: blue >t5, where ρ blue is the apparent reflectivity of the blue light band channel, and t5 is the fifth preset threshold;

[0017] And / or, the third preset condition is: the brightness temperature BT of the long-wave infrared channel infra <t6, where t6 is a sixth preset threshold;

[0018] And / or, the third preset condition is: And blue >t8, where is the standard deviation of the apparent reflectance of the blue light channel within a radius of 3*3, t7 is the seventh preset threshold, and t8 is the eighth preset threshold;

[0019] And / or, the third preset condition is: in, is the standard deviation of the apparent reflectance of the second near-infrared channel within a radius of 3*3, and t9 is the ninth preset threshold.

[0020] Further, determining whether the preprocessed data satisfies a fourth preset condition, and if so, marking the pixel as a clear sky pixel includes:

[0021] The fourth preset condition is: blue <t10, if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t10 is the tenth preset threshold, and t10<t5;

[0022] And / or, the fourth preset condition is: nir1 -ρ nir3 <t14, if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t14 is the fourteenth threshold;

[0023] And / or, the fourth preset condition is: And t17<ρ blue <t18, if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t15, t16, t17 and t18 are the fifteenth preset threshold, the sixteenth preset threshold, the seventeenth preset threshold and the eighteenth preset threshold respectively;

[0024] Or, the fourth preset condition is: t11<ρ blue <t12 and DEM> t13, if the judgment is yes, the pixel is marked as a brighter clear sky pixel, where t11 and t12 are the eleventh preset threshold and the twelfth preset threshold respectively, DEM is the altitude, the unit is meter, t13 is the preset threshold of the brighter surface altitude, and t10≤t11 <t12≤t5。

[0025] Further, the performance reflectivity data ρ λ The pre-processing includes: using water vapor, ozone and carbon dioxide to perform gas absorption correction respectively to obtain the apparent reflectivity after gas absorption correction.

[0026] Further, the first preset threshold t1: t1=0.1, the second preset threshold t2: t2=285, the third preset threshold t3: t3=0.1, the fourth preset threshold t4: t4=0.08, the fifth preset threshold t5: t5=0.45, the sixth preset threshold t6: t6=250, the seventh preset threshold t7: t7=0.0075, the eighth preset threshold t8: t8=0.4, the ninth preset threshold t9: t9=0.0025.

[0027] Further, the tenth preset threshold t10: t10=0.2, the eleventh preset threshold t11: t11=0.2, the twelfth preset threshold t12: t12=0.25, the thirteenth preset threshold t13: t13=150, the fourteenth preset threshold t14: t14=0, the fifteenth preset threshold t15: t15=0.5, the sixteenth preset threshold t16: t16=0.65, the seventeenth preset threshold t17: t17=0.25, the eighteenth preset threshold t18: t18=0.3.

[0028] The technical solution provided by the embodiment of the present invention brings at least the following beneficial technical effects:

[0029] The present invention provides a haze identification method combining satellite remote sensing multispectral and geographic information. Aiming at the problem that the existing cloud and haze identification algorithm has poor identification accuracy in heavy pollution weather, the reflectance data and brightness temperature data of multispectral remote sensing satellite image pixels are used to identify the range of haze. Since haze is mainly distributed in plains and valleys with low altitudes, and clouds are mainly generated at high altitudes, it is more effective and scientific to identify the range of clear sky and haze based on the altitude auxiliary threshold method. The method makes the selection of threshold easier, and the boundary judgment of clear sky and haze area clearer. For the post-monitoring analysis of special pollution events, the identification conditions can be changed by adjusting the threshold to achieve the best application effect, and the universality is stronger. At the same time, the method improves the accuracy of different pixel identification, so that the detection of clouds, clear sky and haze can be realized quickly and stably, the accuracy of haze identification is improved, and the necessary basis is provided for further research on the interaction between clouds and aerosols, etc., and the user experience is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of a haze identification method combining satellite remote sensing multi-spectral and geographic information provided in the first embodiment of the present invention;

[0031] Figure 2 This is a flow chart of another haze identification method combining satellite remote sensing multi-spectral and geographic information provided in the first embodiment of the present invention;

[0032] Figure 3 This is a flow chart of a haze identification method combining satellite remote sensing multi-spectral and geographic information provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0033] It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The terms "first", "second", etc. in the specification, claims and drawings of the present invention are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, such as processes, methods, systems, products or devices that include a series of steps S or units are not necessarily limited to those steps S or units that are clearly listed, but may include other steps S and units that are not clearly listed or inherent to these processes, methods, products or devices.

[0034] In order to enable those skilled in the art to better understand the scheme of the present invention, the scheme in the embodiment of the present invention is clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0035] Embodiment 1

[0036] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0037] Figure 1 The present embodiment provides a method for identifying haze using elevation-assisted cloud and haze, including:

[0038] Step S10: Obtaining reflectance data of at least two preset spectral channels of the remote sensing satellite image λ and brightness temperature data BT, and preprocessing the data to obtain preprocessed data, wherein the at least two preset spectral channels respectively use a visible light channel and an infrared channel.

[0039] The apparent reflectivity refers to the apparent reflectivity of the top of the atmosphere, which is the sum of the surface reflectivity and the atmospheric reflectivity.

[0040] During the specific implementation process, the satellite observation primary data is read for each pixel in at least two preset spectral channels, for example, the image grayscale value DN of the visible light, near infrared, shortwave infrared and thermal infrared channels are read respectively, and then the count value is converted into the apparent reflectivity and brightness temperature of the top of the atmosphere through the radiation calibration coefficient and the physical quantity conversion formula.

[0041] Among them, the physical quantity conversion formula includes:

[0042] Among them, ρ λ is the reflectivity, d is the distance between the sun and the earth, ESUN λ is the solar irradiance, θ is the solar altitude angle;

[0043] And, L λ =Gain*DN+Offset...(2), where L λ is the equivalent radiance at the entrance pupil of the satellite payload channel, in W / (m 2 *sr*μm), Gain and Offset are calibration coefficient gain offset, unit W / (m 2 *sr*μm);

[0044] as well as, Where BT is the brightness temperature at the top of the atmosphere, and K1 and K2 are conversion constants.

[0045] Step S20, performing the following judgment on each pixel in the remote sensing satellite image: judging whether the preprocessed data meets a first preset condition, and if so, marking the pixel as an ice and snow pixel.

[0046] Specifically, since ice and snow pixels are a type of cloud pixels, they have a stronger reflectivity than ordinary cloud pixels, and their atmospheric reflectivity and apparent reflectivity are the strongest among all pixels to be tested. Therefore, it is easier to identify ice and snow pixels through the apparent reflectivity of one or several preset spectral channels.

[0047] In a preferred exemplary embodiment, whether the preprocessed data satisfies the first preset condition is that the reflectance of the pixel to be measured in the near-infrared channel is greater than or equal to the preset reflectance threshold, for example, the preset reflectance threshold is 0.8, or whether the preprocessed data satisfies the first preset condition is that the brightness temperature BT of the pixel to be measured in the near-infrared channel is greater than or equal to the preset brightness temperature threshold, for example, the brightness temperature threshold is BT 1 , then BT ≥ BT 1 .

[0048] Step S30: If the result of step S20 is no, then determine whether the preprocessed data meets a second preset condition; if the result of step S20 is yes, then mark the pixel as an inland water pixel.

[0049] Specifically, as a type of cloud pixel, the inland water pixel has a lower reflectivity than ordinary cloud pixels, and its atmospheric reflectivity and apparent reflectivity are the lowest among all pixels to be tested. Therefore, it is easier to identify inland water pixels through the apparent reflectivity of one or several preset spectral channels.

[0050] In a preferred exemplary embodiment, whether the preprocessed data satisfies the second preset condition is that the reflectance of the pixel to be measured in the near-infrared channel is less than or equal to the preset reflectance threshold, for example, the preset reflectance threshold is 0.2, or whether the preprocessed data satisfies the first preset condition is that the brightness temperature BT of the pixel to be measured in the near-infrared channel is less than or equal to the preset brightness temperature threshold, for example, the brightness temperature threshold is BT 2 , then BT≤BT 2 .

[0051] Step S40: If the result of step S30 is no, then determine whether the preprocessed data meets a third preset condition; if the result of step S30 is yes, then mark the pixel as a cloud pixel.

[0052] Specifically, since the reflectivity and brightness temperature of cloud pixels are quite different from those of ice and snow pixels and inland water pixels, and their value range is between the two, it is necessary to make a specific distinction between the types of cloud pixels to improve the recognition accuracy of cloud and haze.

[0053] In a preferred exemplary embodiment, whether the preprocessed data satisfies the third preset condition is that the reflectance of the pixel to be measured in the near-infrared channel is within the preset reflectance threshold range, for example, the preset reflectance threshold value range is [0.4, 0.6], or whether the preprocessed data satisfies the first preset condition is that the brightness temperature of the pixel to be measured in the visible near-infrared channel is within the preset brightness temperature threshold range, for example, the brightness temperature threshold value range is [BT 2 ,BT 1 ].

[0054] Step S50. If the judgment result of step S40 is no, then determine whether the preprocessed data meets the fourth preset condition. If the judgment result is yes, then mark the pixel as a clear sky pixel. If the judgment result is no, then mark the pixel as a haze pixel. The fourth preset condition includes an altitude geographic information preset condition and a spectral change rate preset condition.

[0055] Specifically, the apparent reflectance and brightness temperature of the clear sky pixel are significantly different from those of the ice and snow pixel, the inland water pixel, and the cloud pixel. λ When the brightness temperature data BT meets the fourth preset condition, the pixel is judged to be a clear sky pixel.

[0056] Among them, the altitude geographic information preset condition, that is, the threshold is set according to the altitude change information, so as to exclude the measured pixel types at different altitudes. Since haze is mainly distributed in plains and valleys with low altitudes, while clouds are mainly generated at high altitudes, and due to the geographical location such as altitude and landform pressure difference, the apparent reflectivity and brightness temperature of clear sky pixels are quite different, so the altitude geographic information preset condition can effectively identify a large range of cloud pixels and clear sky pixels, thereby improving the recognition accuracy of haze.

[0057] Among them, the preset conditions of the spectral change rate are to use spectral channels of different wavelength ranges to set multi-condition thresholds respectively, and the thresholds of the preset conditions of the optimal spectral change rate are inverted by methods such as tree analysis. For example, data reading and calculation are performed on the measured pixels through visible light channels of different wavelengths and infrared channels of different wavelengths, and then the data is compared with the optimal threshold, and the attribute range of the measured pixels is determined by the elimination method.

[0058] It should be noted that the first, second, third and fourth preset conditions may be the performance reflectance data ρ in multiple preset spectral channels. λ Function combination of brightness temperature data BT, such as the reflectance data ρ of the red band channel and the near infrared channel λ The ratio of the difference between the difference and the sum is greater than the preset threshold value of the reflectivity, and the brightness temperature data BT is less than the preset threshold value of the brightness temperature. It can also be other possible function combinations, which are not limited here. In addition, the satellite data used in the present invention meets the bands defined by the above terms, and the data has the latitude and longitude values ​​of the pixels.

[0059] In the embodiment of the present invention, since haze is mainly distributed in plains and valleys with lower altitudes, and clouds are mainly generated at high altitudes, the elevation data-assisted threshold method is mainly used for cloud and haze identification in heavily polluted areas with large terrain undulations, which can improve the accuracy of cloud and haze identification.

[0060] It can be seen that in the embodiment of the present invention, the haze identification method combining satellite remote sensing multi-spectral and geographic information has at least the following technical effects compared with the prior art: it makes the selection of thresholds easier, the boundary judgment of clear sky and haze areas clearer, and for the post-monitoring analysis of special pollution events, the identification conditions can be changed by adjusting the thresholds to achieve the best application effect, and the universality is stronger. At the same time, the accuracy of different pixel identification is improved, so that the detection of clouds, clear sky pixels and haze can be realized quickly and stably, the accuracy of haze identification is improved, and the necessary basis is provided for further research on the interaction between clouds and aerosols, and the user experience is improved.

[0061] In a preferred embodiment, the preset spectral channels include: a blue light band channel, a red light band channel, a first near infrared channel, a second near infrared channel, a third near infrared channel, a fourth near infrared channel and a long wave infrared channel;

[0062] The spectral band center wavelength λ of the preset spectral channel is respectively expressed as: blue light band channel λ blue , red light band channel λ red , the first near-infrared channel λ nir1 , the second near-infrared channel λ nir2 , the third near infrared channel λ nir3 , the fourth near-infrared channel λ nir4 and the long-wave infrared channel λ infra , where 0.8μm≤λ nir1 ≤0.9μm, 1.3μm≤λ nir2 ≤1.4μm, 1.6μm≤λ nir3 ≤1.7μm, 2.1μm<λ nir4 ≤2.3μm, 7μm≤λ infra ≤14μm.

[0063] Specifically, a multispectral remote sensing satellite sensor usually has a visible light channel and an infrared channel. According to the different central wavelength ranges, the visible light channel is divided into a blue light band channel, a red light band channel, a green light band channel, etc., and the infrared channel can be divided into a near infrared channel, a medium-wave infrared channel, and a long-wave infrared channel, etc. In the embodiment of the present invention, the central wavelengths of the first, second, third, and fourth near-infrared channels are all located in the near-infrared band of 0.8 μm to 2.3 μm, and the long-wave infrared channel is located in the long-wave infrared band of 7 μm to 14 μm.

[0064] In a preferred embodiment, the central wavelength λ of the first near-infrared channel nir1 is 0.865 μm, the central wavelength λ of the second near-infrared channel nir2 is 1.38 μm, the central wavelength λ of the third near-infrared channel nir3 is 1.64 μm, the central wavelength λ of the fourth near-infrared channel nir4 The central wavelength of the long-wave infrared channel is 2.13 μm and λ infra It is 11μm.

[0065] Since the preset spectral channels in each band range have different reflectivity for the measured pixels, and due to the influence of factors such as the solar radiation period and weather, some bands will be filtered out when the measured pixels are obtained in a single preset spectral channel, resulting in distortion of the measured values. Therefore, the measured pixels are photographed and read simultaneously through multiple preset spectral channels, and then jointly analyzed and according to certain analysis methods, such as decision tree analysis, the optimal function combination of multiple preset spectral channels is inverted, thereby improving the recognition accuracy of cloud and haze.

[0066] In the embodiment of the present invention, the multiple preset spectral channels cover the most commonly used visible light bands and infrared bands, which can ensure that data such as the reflectivity and brightness temperature of the measured pixels in each band can be effectively read, thereby maximizing the accuracy of cloud and haze identification.

[0067] It should be noted that, in other embodiments, the preset spectral channel may adopt other preset spectral channels, such as a combination of a green light band channel and a near-infrared channel, and the embodiment of the present invention does not limit this.

[0068] In a preferred embodiment, the first preset condition is: And BT infra <t2, where ρ nir1 is the apparent reflectance of the first near-infrared channel, ρ nir3 is the apparent reflectance of the third near-infrared channel, t1 is the first preset threshold, BT infra is the brightness temperature of the long-wave infrared channel, and t2 is the second preset threshold.

[0069] Specifically, the ratio of the difference between the apparent reflectance of the first near-infrared channel in the 0.8μm to 0.9μm band and the third near-infrared channel in the 1.6μm to 1.7μm band and the sum of the two is greater than a first preset threshold, wherein the first preset threshold is the threshold for the ratio of the two to be identified as ice and snow. At the same time, the brightness temperature BT of the long-wave infrared channel that meets the condition infra If the pixel is smaller than a second preset threshold, the pixel can be marked as an ice and snow pixel, wherein the second preset threshold is the threshold for identifying the band as ice and snow.

[0070] During the specific implementation process, the first and second preset thresholds are set according to specific conditions such as satellite sensor parameters, seasons, etc. The first and second preset thresholds are obtained by downloading and inverting data, which can be used as the first and second preset thresholds for a period of time in the measured area.

[0071] In the embodiment of the present invention, the first preset condition includes the reflectance data of two near-infrared channels of different bands and the brightness temperature data of a long-wave infrared channel, which effectively prevents the problem of data distortion. The steps are simple and easy to operate, and the accuracy of cloud and haze identification is improved.

[0072] In a preferred embodiment, the second preset condition is: And nir4 <t4, where ρ red are the apparent reflectances of the red light band channel respectively, and t3 is the third preset threshold; nir4 is the apparent reflectivity of the fourth near-infrared channel, and t4 is the fourth preset threshold.

[0073] Specifically, the third preset threshold and the fourth preset threshold are thresholds for determining inland water pixels. The specific implementation process may refer to the first and second preset thresholds in the above-mentioned embodiment of the present invention, and will not be repeated here.

[0074] In an embodiment of the present invention, the second preset condition includes reflectance data of a red light band channel, a near infrared channel and a long-wave infrared channel, covering a wider band range, which can effectively prevent data distortion problems, and the steps are simple and easy to operate, thereby improving the accuracy of cloud and haze identification.

[0075] In a preferred embodiment, the third preset condition is: blue >t5, where ρ blue is the apparent reflectivity of the blue light band channel, and t5 is the fifth preset threshold;

[0076] And / or, the third preset condition is: the brightness temperature BT of the long-wave infrared channel infra <t6, where t6 is a sixth preset threshold;

[0077] And / or, the third preset condition is: And blue >t8, where is the standard deviation of the apparent reflectance of the blue light channel within a radius of 3*3, t7 is the seventh preset threshold, and t8 is the eighth preset threshold;

[0078] And / or, the third preset condition is: in, is the standard deviation of the apparent reflectance of the second near-infrared channel within a radius of 3*3, and t9 is the ninth preset threshold.

[0079] It should be noted that the third preset condition described in the embodiment of the present invention includes multiple "and / or" parallel conditions, that is, if one or at least two of the conditions in the embodiment of the present invention are met, it can be judged as yes and the measured pixel is marked as a cloud pixel.

[0080] Specifically, due to the wide distribution and many types of cloud pixels, for example, intermediate clouds include cirrus, stratus, altocumulus and other forms, the apparent reflectance data and brightness temperature data involved are quite different and cover a wide range. Therefore, it is necessary to perform multi-spectral channel collection and identification under multiple conditions for intermediate cloud pixels in different situations.

[0081] In the embodiment of the present invention, the fifth to ninth preset thresholds are all thresholds for determining cloud pixels. For specific implementations, reference may be made to the detailed description of the above embodiment of the present invention, which will not be repeated here.

[0082] In an embodiment of the present invention, the third preset condition includes the apparent reflectance data of a blue light band channel and a near-infrared channel and the brightness temperature data of a long-wave infrared channel, and also includes the standard deviation of the apparent reflectance of the blue light channel within a 3*3 radius and the standard deviation of the apparent reflectance of the second near-infrared channel within a 3*3 radius for threshold judgment, where the unit of the radius is m, and the covered band range is wider, which can effectively prevent data distortion problems, reduce accidental errors, and improve the accuracy of cloud and haze identification.

[0083] Figure 2 FIG. 1 is a flow chart of another haze identification method combining satellite remote sensing multi-spectral and geographic information provided by an embodiment of the present invention. Figure 2 As shown, in a preferred embodiment, judging whether the preprocessed data satisfies the fourth preset condition, and if the judgment is yes, marking the pixel as a clear sky pixel includes:

[0084] The fourth preset condition is: blue <t10, if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t10 is the tenth preset threshold, and t10<t5;

[0085] And / or, the fourth preset condition is: nir1 -ρ nir3 <t14, if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t14 is the fourteenth threshold;

[0086] And / or, the fourth preset condition is: And t17<ρ blue <t18, if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t15, t16, t17 and t18 are the fifteenth preset threshold, the sixteenth preset threshold, the seventeenth preset threshold and the eighteenth preset threshold respectively;

[0087] Or, the fourth preset condition is: t11<ρ blue<t12 and DEM>t13, if the judgment is yes, the pixel is marked as a brighter clear sky pixel, wherein t11 and t12 are the eleventh preset threshold and the twelfth preset threshold respectively, DEM is the altitude, the unit is meter, t13 is the preset threshold for the brighter surface altitude, and t10≤t11<t12≤t5.

[0088] It should be noted that the third preset condition in the embodiment of the present invention includes multiple "and / or" parallel conditions, that is, if one or at least two of the conditions in the embodiment of the present invention are met, it can be judged as yes, and the measured pixel is marked as a clear sky pixel. At the same time, the third preset condition also includes an "or" condition, that is, if at least one of the conditions before and after the "or" is met, it can be judged as yes.

[0089] Among them, due to the difference in pressure due to geographical location, such as altitude and topography, the apparent reflectance and brightness temperature of clear sky pixels vary greatly. Therefore, it is necessary to specifically distinguish the types of clear sky pixels to improve the recognition accuracy of clouds and haze. According to the brightness of clear sky pixels, they can be divided into bright clear sky pixels and brighter clear sky pixels. At the same time, due to the difference in altitude, as well as the influence of factors such as snow and vegetation, the apparent reflectance and brightness temperature of the brighter clear sky pixel area have a large differential pressure. Therefore, it is necessary to introduce the judgment basis of the altitude threshold to improve the accuracy of cloud and haze recognition.

[0090] Specifically, the tenth to eighteenth preset thresholds are all thresholds for distinguishing clear sky pixels. The specific implementation method can refer to the detailed description of the above-mentioned embodiment of the present invention, which will not be repeated here.

[0091] In an embodiment of the present invention, the fourth preset condition includes the reflectance of a blue light band channel, a red light band channel, and three near-infrared channels. At the same time, an altitude threshold setting is adopted when identifying brighter clear sky pixels. The covered band range is wider, which can effectively prevent data distortion problems, reduce accidental errors, and improve the accuracy of cloud and haze identification.

[0092] In a preferred embodiment, the performance reflectivity data ρ λ The preprocessing includes performing gas absorption correction on the apparent reflectance of a preset spectral channel to obtain the gas absorption corrected apparent reflectance.

[0093] Preferably, performing gas absorption correction on the apparent reflectance of a preset spectral channel includes: performing gas absorption correction using water vapor, ozone, and carbon dioxide, respectively.

[0094] Specifically, since the remote sensing accuracy varies in different gas aerosol conditions and the gas composition and content of the air vary greatly due to factors such as climate change, the accuracy of images acquired by remote sensing satellites cannot be guaranteed. Therefore, it is necessary to perform gas absorption correction using several commonly used gases.

[0095] According to research and experiments, the accuracy of remote sensing is greatly improved by correcting the gas absorption error. Zhang Junhua, Wang Meihua, Mao Jietai, Error Analysis and Correction of Remote Sensing Aerosols with Multi-band Photometer, 2000, 24 (6). P4-P49. The document combines the multi-band photometer used in actual remote sensing aerosols in 1998, and analyzes in detail the errors caused by instrument calibration, instrument field of view, filter width, gas absorption, etc., and proposes corresponding correction methods, especially for water vapor absorption, a three-band differential absorption correction method is proposed. The results show that the accuracy of remote sensing is greatly improved by error correction. The aerosol optical thickness is about 0.1, the relative error does not exceed 10%, and the relative error is about 2% under moderate atmospheric turbidity.

[0096] In the embodiment of the present invention, by using common gases in the air, such as water vapor, ozone and carbon dioxide, to perform gas absorption correction respectively, remote sensing accuracy is improved, accidental errors are reduced, and the accuracy of cloud and haze identification is improved.

[0097] It can be seen that in the embodiment of the present invention, the haze identification method combining satellite remote sensing multi-spectral and geographic information has at least the following technical effects compared with the prior art: it makes the selection of thresholds easier, the boundary judgment of clear sky and haze areas clearer, and for the post-monitoring analysis of special pollution events, the identification conditions can be changed by adjusting the thresholds to achieve the best application effect, and the universality is stronger. At the same time, the accuracy of different pixel identification is improved, so that the detection of clouds, clear sky pixels and haze can be realized quickly and stably, the accuracy of haze identification is improved, and the necessary basis is provided for further research on the interaction between clouds and aerosols, and the user experience is improved.

[0098] Embodiment 2

[0099] Based on the first embodiment, the present invention also provides another haze identification method combining satellite remote sensing multi-spectral and geographic information. Figure 3 A flow chart of a haze identification method combining satellite remote sensing multi-spectral and geographic information is provided in an embodiment of the present invention, such as Figure 3 As shown:

[0100] Step Q10: Read the count values of each pixel in the FY-3D / MERSI-II L1 data in two visible light channels (0.47μm and 0.65μm), three near-infrared channels (0.865μm, 1.38μm, and 1.64μm), one shortwave infrared channel (2.13μm), and one infrared channel (11μm). Through the calibration coefficient and physical quantity conversion formula, convert the calculated values of the above channels into the apparent reflectance or brightness temperature at the top of the atmosphere;

[0101] Step Q20: Use water vapor, ozone, and carbon dioxide to correct the water vapor absorption, ozone absorption, and carbon dioxide absorption for the pixels obtained in the channels of 0.47μm, 0.65μm, 0.865μm, 1.38μm, 1.64μm, and 2.13μm bands respectively, and obtain the apparent reflectance after gas absorption correction;

[0102] Step Q30: For each pixel, when t1 = 0.1 and t2 = 285, make the following judgment for each pixel in the remote sensing satellite image: Judge whether the preprocessed data meets the following judgment conditions: and BT infra < t2. If the judgment is yes, mark the pixel as an ice and snow pixel; otherwise, continue with the following process;

[0103] Step Q40: For each non-ice and snow pixel, when t3 = 0.1 and t4 = 0.08, then judge whether the preprocessed data meets the following judgment conditions: and ρ nir4 < t4. If the judgment is yes, mark the pixel as an inland water body pixel; otherwise, continue with the following process;

[0104] Step Q50: For each pixel that is neither ice and snow nor an inland water body, when t5 = 0.45, execute the following judgment condition: ρ blue > t5. If it is satisfied, mark the pixel as a cloud pixel;

[0105] Step Q60: For each pixel that is neither ice and snow nor an inland water body, when t6 = 250, execute the following judgment condition: BT infra < t6. If it is satisfied, mark the pixel as a cloud pixel;

[0106] Step Q70: For each pixel that is neither ice and snow nor an inland water body, when t7 = 0.0075 and t8 = 0.4, execute the following judgment condition: and ρ blue > t8. If it is satisfied, mark the pixel as a cloud pixel;

[0107] Step Q80: For each pixel that is neither ice nor snow nor inland water, when t9=0.0025, the following judgment conditions are executed: in, is the standard deviation of the apparent reflectance of the blue light channel within a 3*3 radius. If it satisfies the standard deviation, the pixel is marked as a cloud pixel.

[0108] Step Q90: For each pixel that is not ice or snow, not inland water, and not cloud, when t10 = 0.2, execute the following judgment condition: blue <t10, where is the standard deviation of the apparent reflectance of the second near-infrared channel within a 3*3 radius. If it meets the standard deviation, the pixel is marked as a clear sky pixel.

[0109] Step Q100: For each pixel that is not ice and snow, not inland water, and not cloud, when t11 = 0.2, t12 = 0.25, and t13 = 150, the following judgment condition is executed: t11 < ρ blue <t12 and DEM>t13, if satisfied, the pixel is marked as a clear sky pixel;

[0110] Step Q110: For each pixel that is not ice or snow, not inland water, and not cloud, when t14=0, execute the following judgment condition: nir1 -ρ nir3 <t14, if satisfied, the pixel is marked as a clear sky pixel;

[0111] Step Q120: For each pixel that is not ice and snow, not inland water, and not cloud, when t15=0.5, t16=0.65, t17=0.25, and t18=0.3, the following judgment conditions are performed: And t17<ρ blue <t18, if it is satisfied, the pixel is marked as a clear sky pixel, and all pixels of the unmarked category are marked as haze.

[0112] The embodiment of the present invention is a preferred exemplary embodiment. The value of each preset threshold value depends on the conditions such as climate, satellite sensor parameters and other factors, and the preferred value is calculated by the inversion method. The comparison of its cloud and haze recognition results with real-time cloud and haze data shows that the result of this method has high accuracy.

[0113] It can be seen that in the embodiment of the present invention, the haze identification method combining satellite remote sensing multi-spectral and geographic information has at least the following technical effects compared with the prior art: it makes the selection of thresholds easier, the boundary judgment of clear sky and haze areas clearer, and for the post-monitoring analysis of special pollution events, the identification conditions can be changed by adjusting the thresholds to achieve the best application effect, and the universality is stronger. At the same time, the accuracy of different pixel identification is improved, so that the detection of clouds, clear sky pixels and haze can be realized quickly and stably, the accuracy of haze identification is improved, and the necessary basis is provided for further research on the interaction between clouds and aerosols, and the user experience is improved.

[0114] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A haze identification method combining satellite remote sensing multispectral and geographic information, It is characterized in that include: Step 1: Obtain reflectance data of at least two preset spectral channels of remote sensing satellite images and brightness temperature data BT, and preprocessing the data to obtain preprocessed data, wherein the at least two preset spectral channels are respectively a visible light channel and an infrared channel; Step 2: for each pixel in the remote sensing satellite image, the following judgment is made: whether the preprocessed data meets a first preset condition, and if so, the pixel is marked as an ice and snow pixel; Step 3: If the result of step 2 is no, then determine whether the preprocessed data meets the second preset condition; if the result is yes, then mark the pixel as an inland water pixel; Step 4: If the result of step 3 is no, then determine whether the preprocessed data meets a third preset condition; if the result of step 3 is yes, then mark the pixel as a cloud pixel; Step 5: If the result of step 4 is no, then determine whether the preprocessed data meets the fourth preset condition; if the result is yes, then mark the pixel as a clear sky pixel; if the result is no, then mark the pixel as a haze pixel, wherein the fourth preset condition includes an altitude geographic information preset condition and a spectral change rate preset condition; in, …(1); is the reflectivity, d is the distance between the sun and the earth, is the solar irradiance, is the solar altitude angle, …(2)， is the equivalent radiance at the entrance pupil of the satellite payload channel, in W / (m 2 *sr*μm), and They are the calibration coefficient, gain and offset, in W / (m 2 *sr*μm), DN is the image gray value; …(3), where BT is the brightness temperature at the top of the atmosphere, K1 and K2 are conversion constants; The preset spectral channels include: a blue light band channel, a red light band channel, a first near infrared channel, a second near infrared channel, a third near infrared channel, a fourth near infrared channel and a long-wave infrared channel; the spectral band center wavelength of the preset spectral channels Respectively represented as: blue light band channel , Red light band channel , the first near-infrared channel , the second near infrared channel , the third near infrared channel , the fourth near infrared channel and long-wave infrared channels , where 0.8µm≤ ≤0.9µm, 1.3µm≤ ≤1.4µm, 1.6µm≤ ≤1.7µm, 2.1µm< ≤2.3µm, 7µm≤ ≤14µm; The first preset condition is: and ,in, is the apparent reflectance of the first near-infrared channel, is the apparent reflectance of the third near-infrared channel, t1 is the first preset threshold, is the brightness temperature of the long-wave infrared channel, and t2 is the second preset threshold; The second preset condition is: and ,in, are respectively the apparent reflectances of the red light band channel, and t3 is the third preset threshold; is the apparent reflectance of the fourth near-infrared channel, and t4 is a fourth preset threshold; The third preset condition is: ,in, is the apparent reflectivity of the blue light band channel, and t5 is the fifth preset threshold; And / or, the third preset condition is: the brightness temperature of the long-wave infrared channel , wherein t6 is the sixth preset threshold; And / or, the third preset condition is: and ,in, is the standard deviation of the apparent reflectance of the blue light channel within a radius of 3*3, t7 is the seventh preset threshold, and t8 is the eighth preset threshold; And / or, the third preset condition is: ,in, is the standard deviation of the apparent reflectance of the second near-infrared channel within a radius of 3*3, and t9 is the ninth preset threshold; The fourth preset condition is: , if the judgment is yes, then the pixel is marked as a bright clear sky pixel, where t10 is the tenth preset threshold, and ; And / or, the fourth preset condition is: , if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t14 is the fourteenth threshold; And / or, the fourth preset condition is: and , if the judgment is yes, the pixel is marked as a bright clear sky pixel, wherein t15, t16, t17 and t18 are the fifteenth preset threshold, the sixteenth preset threshold, the seventeenth preset threshold and the eighteenth preset threshold respectively; Or, the fourth preset condition is: and , if the judgment is yes, the pixel is marked as a brighter clear sky pixel, wherein t11 and t12 are the eleventh preset threshold and the twelfth preset threshold respectively, DEM is the altitude in meters, t13 is the preset threshold for the brighter surface altitude, and ; The first preset threshold t1=0.1, the second preset threshold t2=285, the third preset threshold t3=0.1, the fourth preset threshold t4=0.08, the fifth preset threshold t5=0.45, the sixth preset threshold t6=250, the seventh preset threshold t7=0.0075, the eighth preset threshold t8=0.4, the ninth preset threshold t9=0.0025, the tenth preset threshold t10=0.2, the eleventh preset threshold t11=0.2, the twelfth preset threshold t12=0.25, the thirteenth preset threshold t13=150, the fourteenth preset threshold t14=0, the fifteenth preset threshold t15=0.5, the sixteenth preset threshold t16=0.65, the seventeenth preset threshold t17=0.25, and the eighteenth preset threshold t18=0.

3.

2. A haze identification method combining satellite remote sensing multispectral and geographic information as claimed in claim 1, It is characterized in that The central wavelength of the first near-infrared channel The central wavelength of the second near-infrared channel is 0.865µm. The central wavelength of the third near-infrared channel is 1.38µm. The central wavelength of the fourth near-infrared channel is 1.64µm. The central wavelength of the 2.13µm and long-wave infrared channels 11µm.

3. The haze identification method combining satellite remote sensing multi-spectral and geographic information as claimed in claim 1, It is characterized in that Reflectivity data for the performance The pre-processing includes: using water vapor, ozone and carbon dioxide to perform gas absorption correction respectively to obtain the apparent reflectivity after gas absorption correction.

Images