A high-precision cloud phase identification method based on data fusion

Abstract

The present invention discloses a high-precision cloud phase recognition method based on data fusion. The method comprises obtaining the full-disk 4km L1-level data of the Fengyun-4B imager, the 1km L1-level data of the Fengyun-3E medium-resolution spectral imager and its positioning data, and extracting the data of the corresponding radiation wavelength; calibrating the extracted data after parsing, and calculating the corresponding radiance, and converting the radiance into infrared channel brightness temperature data; parsing the infrared channel temperature measurement data of the Fengyun-4B as the bottom layer, and fusing the existing infrared channel temperature measurement data of the Fengyun-3E with the bottom layer; then detecting the presence or absence of clouds in the target area, so as to divide the target area into target pixels with clouds and target pixels without clouds, and then performing cloud phase recognition. The present invention can effectively identify ice cloud areas, warm water cloud areas, supercooled water clouds and mixed water clouds, greatly improving the accuracy of cloud phase recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing and meteorological observation, and in particular to a high-precision cloud phase recognition method based on data fusion. Background Art

[0002] With the advancement of satellite remote sensing technology and the continuous development of the Fengyun series of satellites, the remote sensing data obtained by Fengyun-4B and Fengyun-3E have provided important support for meteorological research and weather forecasting. Fengyun-4B mainly obtains time series data of clouds through geosynchronous orbit. Its advantage lies in the high-frequency monitoring of real-time meteorological changes, but it may be limited in spatial resolution and observation capabilities in high-latitude areas. Fengyun-3E satellites use polar orbits and can provide high-resolution and fine-grained observation data on a global scale.

[0003] Existing cloud phase identification has the following disadvantages:

[0004] 1. Insufficient cloud phase recognition accuracy: The cloud phase recognition accuracy in existing technologies is limited by fixed thresholds and simplified models, which makes it difficult to adapt to complex and changing meteorological conditions;

[0005] 2. Low data processing efficiency: Traditional methods involve complex data analysis and processing, and heavy computational burden;

[0006] 3. Limited scope of application: The existing system is not capable of identifying clouds in extreme weather or high-latitude areas;

[0007] 4. Poor real-time performance: Existing systems may find it difficult to provide high-frequency real-time updates. Summary of the invention

[0008] The purpose of the present invention is to provide a high-precision cloud phase recognition method based on data fusion in view of the deficiencies in the prior art.

[0009] To achieve the above object, the present invention provides a high-precision cloud phase recognition method based on data fusion, comprising:

[0010] Obtain the 4km L1 data of the full disk of the FY-4B imager, the 1km L1 data of the FY-3E medium-resolution spectral imager and their positioning data, and extract data with radiation wavelengths of 0.7 microns, 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns;

[0011] The extracted data is parsed, and then the parsed radiation wavelengths of 0.7 microns, 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns are calibrated in combination with the infrared channel calibration coefficient, and the corresponding radiance is calculated and converted into infrared channel brightness temperature data;

[0012] The infrared channel temperature data of FY-4B is analyzed as the bottom layer. When the data of FY-3E exists within the observation range, the infrared channel brightness temperature data of FY-3E is analyzed, and the infrared channel temperature data of FY-4B is matched with the infrared channel temperature data of FY-3E in time and space. Then, the infrared channel temperature data of FY-3E is fused with the bottom layer.

[0013] Using the reflectivity data with a radiation wavelength of 0.7 microns and the brightness temperature data with radiation wavelengths of 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns, the target area is detected for cloudiness, so as to divide the target area into target pixels with clouds and target pixels without clouds, and the target pixels without clouds are identified as clear sky areas;

[0014] Pixels with a radiation wavelength of 10.8 microns and a cloud top temperature less than 233K in target pixels with clouds are identified as ice cloud areas, and pixels with a radiation wavelength of 10.8 microns and a cloud top temperature greater than 273K in target pixels with clouds are identified as warm water cloud areas. If the cloud top temperature with a radiation wavelength of 10.8 microns in target pixels with clouds is between 233K and 273K, determine whether the temperature difference between the radiation wavelengths of 7.42 microns and 10.8 microns meets the following conditions:

[0015]

[0016] Among them, BTD is the brightness temperature difference between the radiation wavelengths of 7.42 microns and 10.8 microns, and CTT is the brightness temperature at the radiation wavelength of 10.8 microns. If it meets the requirements, it is identified as a supercooled water cloud, otherwise it is identified as a mixed water cloud.

[0017] Furthermore, the full-disk 4kmL level 1 data of the FY-4B imager is acquired every 15 minutes.

[0018] Furthermore, the L1 level data of the medium-resolution spectral imager of the Fengyun-3E satellite and its positioning data are acquired every 5 minutes.

[0019] Furthermore, it also includes:

[0020] The identification results are generated into cloud phase identification products in the 4th category format of the meteorological information comprehensive analysis and processing system and archived for display on the meteorological information comprehensive analysis and processing system of the business platform.

[0021] Furthermore, the calibration method is specifically as follows:

[0022]

[0023] Wherein, λ is the radiation wavelength, including 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns. The reflectivity obtained by calibrating the data with radiation wavelength λ is: is the slope of the infrared channel calibration coefficient, is the intercept in the infrared channel calibration coefficient, is the data with the resolved radiation wavelength λ.

[0024] Furthermore, the radiance is calculated as follows:

[0025]

[0026] in, is the calculated radiance of the radiation with wavelength λ, is the channel average solar irradiance at the average distance between the sun and the earth, in units of , is the ratio of pi.

[0027] Furthermore, the infrared channel brightness temperature data is converted based on the following formula:

[0028]

[0029] Among them, T is the brightness temperature data of the infrared channel, the unit is K, h is Planck's constant, c is the speed of light, k is the Boltzmann constant, and e is the natural constant.

[0030] Furthermore, the conditions for cloud detection during the day are as follows:

[0031] If the reflectivity of the radiation wavelength of 0.7 microns is higher than the set reflectivity threshold, it is determined to be a target pixel with clouds, otherwise it is determined to be a target pixel without clouds;

[0032] The conditions for cloud detection at night are as follows:

[0033] BTD13.3<260

[0034] BTD[7.42,10.8]=BT7.42-BT10.8>0.85

[0035] BTD[3.75,7.42]=BT3.75-BT7.42>1.4

[0036] BTD[10.8,3.75]=BT10.8-BT3.75>2.3

[0037] BTD[3.75,12]=BT3.75-BT12>2.3

[0038] BTD[12,13.3]=BT12-BT13.3>1.8

[0039] Among them, BTD13.3 is the brightness temperature at a radiation wavelength of 13.3 μm, BTD[7.42, 10.8] is the brightness temperature difference between radiation wavelengths of 13.3 μm and 10.8 μm, BT7.42 is the brightness temperature at a radiation wavelength of 7.42 μm, BT10.8 is the brightness temperature at a radiation wavelength of 10.8 μm, BTD[3.75,7.42] is the brightness temperature difference between radiation wavelengths of 3.75 μm and 7.42 μm, BT3.75 is the brightness temperature at a radiation wavelength of 3.75 μm, BTD[10.8,3.75] is the brightness temperature difference between radiation wavelengths of 10.8 μm and 3.75 μm, BT10.8 is the brightness temperature at a radiation wavelength of 10.8 μm, BTD[3.75,12] is the brightness temperature difference between radiation wavelengths of 3.75 μm and 12 μm, BT12 is the brightness temperature at a radiation wavelength of 12 μm, BTD[12,13.3] is the brightness temperature difference between radiation wavelengths of 12 μm and 13.3 μm, BT13.3 is the brightness temperature at a radiation wavelength of 13.3 μm. If all conditions are met, it is judged as a target pixel with clouds. If one or more conditions are not met, it is judged as a target pixel without clouds.

[0040] Beneficial effects: The present invention effectively integrates the data of FY-4B and FY-3E, and then performs cloud detection to divide the target area into target pixels with clouds and target pixels without clouds. Then, cloud phase recognition is performed on the target pixels with clouds, so as to effectively identify ice cloud areas, warm water cloud areas, supercooled water clouds and mixed water clouds, and greatly improve the accuracy of cloud phase recognition. Through more accurate cloud phase recognition, precipitation and extreme weather can be more effectively predicted, thereby supporting better environmental management and disaster warning, indirectly reducing the negative impact on the environment, and promoting the further development and application of meteorological work and MICAPS business platform automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flow chart of a high-precision cloud phase recognition method based on data fusion according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. The embodiments are implemented based on the technical solution of the present invention. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0043] like Figure 1 As shown, an embodiment of the present invention provides a high-precision cloud phase recognition method based on data fusion, comprising:

[0044] The full disk 4km L1 data of the FY-4B imager, the 1km L1 data of the FY-3E medium-resolution spectroscopic imager and their positioning data are obtained, and the target channel data with radiation wavelengths of 0.7 microns, 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, 13.3 microns, etc. are extracted. Since the full disk 4kmL1 data of the FY-4B imager is updated every 15 minutes, and the L1 data of the FY-3E medium-resolution spectroscopic imager and its positioning data are updated every 5 minutes, it is preferred to obtain the full disk 4kmL1 data of the FY-4B imager every 15 minutes, and obtain the L1 data of the FY-3E medium-resolution spectroscopic imager and its positioning data every 5 minutes, so that the acquisition frequency is the same as the data update frequency.

[0045] The extracted radiation wavelengths of 0.7 microns, 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, 13.3 microns and other target channel data are analyzed, and then the analyzed radiation wavelengths of 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns are calibrated using the infrared channel calibration coefficient, and the corresponding radiance is calculated and converted into infrared channel brightness temperature data.

[0046] Specifically, the above-mentioned method of calibrating the analyzed radiation wavelengths of 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns in combination with the infrared channel calibration coefficient is as follows:

[0047]

[0048] Wherein, λ is the radiation wavelength, including 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns. The reflectivity obtained by calibrating the data with radiation wavelength λ is: is the slope of the infrared channel calibration coefficient, is the intercept in the infrared channel calibration coefficient, is the data with the resolved radiation wavelength λ.

[0049] The above radiance is calculated as follows:

[0050]

[0051] in, is the calculated radiance of the radiation with wavelength λ, and its unit is , is the channel average solar irradiance at the average distance between the sun and the earth, in units of , is the ratio of pi.

[0052] The above infrared channel brightness temperature data is converted based on the following formula:

[0053]

[0054] Among them, T is the brightness temperature data of the infrared channel, the unit is K, h is Planck's constant, c is the speed of light, k is the Boltzmann constant, and e is the natural constant.

[0055] The infrared channel temperature data of FY-4B is analyzed as the bottom layer. When the data of FY-3E exists in the observation range, the infrared channel brightness temperature data of FY-3E is analyzed, and the infrared channel temperature data of FY-4B is matched with the infrared channel temperature data of FY-3E in time and space, and then the infrared channel temperature data of FY-3E is merged with the bottom layer. Since the time resolution of the full disk L1 data of the FY-4B imager is 15 minutes, and the time resolution of the L1 data of the medium resolution spectral imager of FY-3E and its positioning data is 5 minutes, the time information data extracted from the full disk L1 data of the FY-4B imager can be matched to the L1 data of the medium resolution spectral imager of FY-3E and its positioning data at the same time, thus completing the above-mentioned time matching processing. Based on the brightness temperature data of the 13th channel (radiation wavelength of 10.8 microns) of the full disk L1 data of the FY-4B imager output by the multi-source data input process, the spatial scale is downscaled to 1000 meters, thus completing the above-mentioned spatial matching processing. After the spatial resolution is unified, the brightness temperature data of the full disk L1 data of the FY-4B imager is used as the bottom layer. When the FY-3E medium resolution spectral imager L1 data exists in the same time observation range, the row and column number information of the brightness temperature array position of the infrared channel extracted from the FY-3E medium resolution spectral imager L1 data in the brightness temperature data of the full disk L1 data of the FY-4B imager is calculated, and the brightness temperature data of the full disk L1 level infrared channel of the FY-4B imager at the corresponding position is replaced with the brightness temperature data of the infrared channel in the FY-3E medium resolution spectral imager L1 data, thus completing the data fusion.

[0056] The reflectivity data with a radiation wavelength of 0.7 microns and the brightness temperature data with radiation wavelengths of 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns are used to detect the presence or absence of clouds in the target area, so as to divide the target area into target pixels with clouds and target pixels without clouds, and identify the target pixels without clouds as clear sky areas.

[0057] Specifically, the cloud detection is divided into daytime and nighttime. Cloud detection during the day is based on the reflectivity of the radiation wavelength of 0.7 microns. Areas with high reflectivity usually indicate that there are clouds or thick clouds, because clouds strongly reflect sunlight. The reflectivity threshold can be set to 0.3, and areas with a reflectivity of 0.7 microns at a radiation wavelength higher than 0.3 are judged as target pixels with clouds, otherwise they are judged as target pixels without clouds.

[0058] Nighttime cloud detection is based on brightness temperature data with radiation wavelengths of 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns. Due to the lack of solar radiation at night, the ground and sensors will not be directly affected by the sun. Part of the long-wave radiation emitted by the surface to the outside world is received by the sensor, and the other part is absorbed or scattered by the atmosphere. If there are clouds above the ground, the surface radiation below the clouds will not be able to completely pass through the clouds and be received by the sensor; the sensor obtains the thermal radiation value of the cloud top, or a mixed value of the cloud top thermal radiation and the surface thermal radiation. Compared with the surrounding underlying surface, the cloud top always has a lower temperature. Based on this temperature difference between the cloud top and the land surface, clouds can be effectively identified using only thermal radiation data. Specifically, the conditions for cloud detection at night are as follows:

[0059] BTD13.3<260

[0060] BTD[7.42,10.8]=BT7.42-BT10.8>0.85

[0061] BTD[3.75,7.42]=BT3.75-BT7.42>1.4

[0062] BTD[10.8,3.75]=BT10.8-BT3.75>2.3

[0063] BTD[3.75,12]=BT3.75-BT12>2.3

[0064] BTD[12,13.3]=BT12-BT13.3>1.8

[0065] Among them, BTD13.3 is the brightness temperature at a radiation wavelength of 13.3 μm, BTD[7.42, 10.8] is the brightness temperature difference between radiation wavelengths of 13.3 μm and 10.8 μm, BT7.42 is the brightness temperature at a radiation wavelength of 7.42 μm, BT10.8 is the brightness temperature at a radiation wavelength of 10.8 μm, BTD[3.75,7.42] is the brightness temperature difference between radiation wavelengths of 3.75 μm and 7.42 μm, BT3.75 is the brightness temperature at a radiation wavelength of 3.75 μm, BTD[10.8,3.75] is the brightness temperature difference between radiation wavelengths of 10.8 μm and 3.75 μm, BT10.8 is the brightness temperature at a radiation wavelength of 10.8 μm, BTD[3.75,12] is the brightness temperature difference between the radiation wavelengths of 3.75 microns and 12 microns, BT12 is the brightness temperature at the radiation wavelength of 12 microns, BTD[12,13.3] is the brightness temperature difference between the radiation wavelengths of 12 microns and 13.3 microns, and BT13.3 is the brightness temperature at the radiation wavelength of 13.3 microns. If all of the above conditions are met, it is judged as a target pixel with clouds. If one or more of the conditions are not met, it is judged as a target pixel without clouds.

[0066] The setting of these parameters is mainly based on the spectral differences in different channels. It mainly includes the spectral differences between the cloud area in the short-wave infrared and mid-wave infrared, which is characterized by BTD[3.7, 7.3]. It also includes the difference between the cloud area in the short-wave infrared and the long-wave infrared, which is characterized by BTD[3.7, 12]. It also includes the spectral differences between different channels in the atmospheric window area of ​​the cloud area, which are characterized by BTD[7.3,12] and BTD[12, 13.9]. Among them, BTD[7.3,12] can be mainly used to monitor thin ice clouds, because the brightness temperature difference of thin ice clouds in the 7.3um and atmospheric window channels is different from that of other ground objects, and it can be used to identify cloud areas.

[0067] If the target pixel is identified as a cloud pixel, the pixel with a radiation wavelength of 10.8 microns and a cloud top temperature less than 233K in the target pixel with clouds is identified as an ice cloud area, and the pixel with a radiation wavelength of 10.8 microns and a cloud top temperature greater than 273K in the target pixel with clouds is identified as a warm water cloud area. If the cloud top temperature with a radiation wavelength of 10.8 microns in the target pixel with clouds is between 233K and 273K, it is determined whether the temperature difference between the radiation wavelengths of 8.55 microns and 10.8 microns meets the following conditions:

[0068]

[0069] Among them, BTD is the brightness temperature difference between the radiation wavelengths of 8.55 microns and 10.8 microns, and CTT is the brightness temperature at the radiation wavelength of 10.8 microns. If it meets the requirements, it is identified as a supercooled water cloud, otherwise it is identified as a mixed water cloud.

[0070] The identification results can also be generated and archived as cloud phase identification products in the Category 4 format of the meteorological information comprehensive analysis and processing system, and can be displayed on the meteorological information comprehensive analysis and processing system of the business platform commonly used by meteorological departments.

[0071] The above is only a preferred embodiment of the present invention. It should be noted that for ordinary technicians in this technical field, other parts not specifically described belong to the prior art or common knowledge. Several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention.

Claims

1. A high-precision cloud phase recognition method based on data fusion, characterized in that: include: Obtain the 4km L1-level data of the full disk imager of FY-4B, the 1km L1-level data of the medium-resolution spectral imager of FY-3E and their positioning data, and extract the data of radiation wavelengths of 0.7 microns, 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns; The extracted data is parsed, and then the parsed radiation wavelengths of 0.7 microns, 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns are calibrated in combination with the infrared channel calibration coefficient, and the corresponding radiance is calculated and converted into infrared channel brightness temperature data; The infrared channel brightness temperature data of FY-4B is analyzed as the bottom layer. When the data of FY-3E exists within the observation range, the infrared channel brightness temperature data of FY-3E is analyzed, and the infrared channel brightness temperature data of FY-4B and FY-3E are matched in time and space. Then, the infrared channel brightness temperature data of FY-3E is fused with the bottom layer. Using the reflectivity data with a radiation wavelength of 0.7 microns and the brightness temperature data with radiation wavelengths of 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns, the target area is detected for cloudiness, so as to divide the target area into target pixels with clouds and target pixels without clouds, and the target pixels without clouds are identified as clear sky areas; Pixels with a radiation wavelength of 10.8 microns and a cloud top temperature less than 233K in target pixels with clouds are identified as ice clouds, and pixels with a radiation wavelength of 10.8 microns and a cloud top temperature greater than 273K in target pixels with clouds are identified as warm water clouds. If the cloud top temperature with a radiation wavelength of 10.8 microns in target pixels with clouds is between 233K and 273K, determine whether the brightness temperature difference between the radiation wavelengths of 7.42 microns and 10.8 microns meets the following conditions: ； Among them, BTD is the brightness temperature difference between the radiation wavelengths of 7.42 microns and 10.8 microns, and CTT is the brightness temperature at the radiation wavelength of 10.8 microns. If it meets the requirements, it is identified as a supercooled water cloud, otherwise it is identified as a mixed water cloud; The conditions for cloud detection during the day are as follows: If the reflectivity of the radiation wavelength of 0.7 microns is higher than the set reflectivity threshold, it is determined to be a target pixel with clouds, otherwise it is determined to be a target pixel without clouds; The conditions for cloud detection at night are as follows: BTD13.3< 260 BTD[7.42,10.8]=BT7.42-BT10.8>0.85 BTD[3.75,7.42]=BT3.75-BT7.42>1.4 BTD[10.8,3.75]=BT10.8-BT3.75>2.3 BTD[3.75,12]=BT3.75-BT12>2.3 BTD[12,13.3]=BT12-BT13.3>1.8 Among them, BTD13.3 is the brightness temperature at a radiation wavelength of 13.3 μm, BTD[7.42, 10.8] is the brightness temperature difference between radiation wavelengths of 7.42 μm and 10.8 μm, BT7.42 is the brightness temperature at a radiation wavelength of 7.42 μm, BT10.8 is the brightness temperature at a radiation wavelength of 10.8 μm, BTD[3.75,7.42] is the brightness temperature difference between radiation wavelengths of 3.75 μm and 7.42 μm, BT3.75 is the brightness temperature at a radiation wavelength of 3.75 μm, BTD[10.8,3.75] is the brightness temperature difference between radiation wavelengths of 10.8 μm and 3.75 μm, BT10.8 is the brightness temperature at a radiation wavelength of 10.8 μm, BTD[3.75,12] is the brightness temperature difference between radiation wavelengths of 3.75 μm and 12 μm, BT12 is the brightness temperature at a radiation wavelength of 12 μm, BTD[12,13.3] is the brightness temperature difference between radiation wavelengths of 12 μm and 13.3 μm, BT13.3 is the brightness temperature at a radiation wavelength of 13.3 μm. If all conditions are met, it is judged as a target pixel with clouds. If one or more conditions are not met, it is judged as a target pixel without clouds.

2. According to the high-precision cloud phase recognition method based on data fusion according to claim 1, it is characterized in that: The full disk 4kmL level 1 data of the Fengyun-4B imager is acquired every 15 minutes.

3. The high-precision cloud phase recognition method based on data fusion according to claim 1 is characterized in that: The L1 level data and positioning data of the medium-resolution spectral imager on the Fengyun-3E satellite are acquired every 5 minutes.

4. The high-precision cloud phase recognition method based on data fusion according to claim 1 is characterized in that: The calibration method is as follows: ； Wherein, λ is the radiation wavelength, including 3.75 microns, 7.42 microns, 10.8 microns, 12 microns, and 13.3 microns. The reflectivity obtained by calibrating the data with radiation wavelength λ is: is the slope of the infrared channel calibration coefficient, is the intercept in the infrared channel calibration coefficient, is the data with the resolved radiation wavelength λ.

5. The high-precision cloud phase recognition method based on data fusion according to claim 4 is characterized in that: The radiance is calculated as follows: ； in, is the calculated radiance of the radiation with wavelength λ, is the channel average solar irradiance at the average distance between the sun and the earth, in units of , is the ratio of pi.

6. A high-precision cloud phase recognition method based on data fusion according to claim 5, characterized in that: The infrared channel brightness temperature data is obtained based on the following formula conversion: ； Among them, T is the brightness temperature data of the infrared channel, the unit is K, h is Planck's constant, c is the speed of light, k is the Boltzmann constant, and e is the natural constant.

Images