A method for monitoring large-scale and long-time series surface exposure rate based on GEE cloud platform

Abstract

The present invention discloses a method for monitoring large-scale and long-time series of surface exposure rate based on GEE cloud platform. Based on the remote sensing big data of GEE cloud platform, through satellite image band information calculation and denoising, a large-scale and long-time series of surface exposure rate monitoring method suitable for land remote sensing is constructed, and the monitoring results can be synchronized to the local area for mapping and analysis, providing data support for regional long-term ecological environment assessment. The present invention solves the problems of small spatial range, short time limit, difficult data integration, serious resource consumption, complicated processing flow, insufficient accuracy, low automation level, etc. in the previous bare surface monitoring process.

Description

Technical Field

[0001] The present invention belongs to the field of satellite remote sensing technology and its application, and in particular to a method for monitoring large-scale and long-time series surface exposure rate based on a GEE cloud platform. Background Art

[0002] Bare surface refers to a phenomenon in which there is no or little vegetation growth and coverage such as forests, shrubs, meadows, etc., resulting in the direct exposure of the ground surface. Due to various reasons such as harsh conditions, climate interference, soil erosion, and human activities, there are two types of bare surface: one is "primary bare surface", that is, there is no vegetation cover for a long period of time or the vegetation has been completely destroyed, such as natural wasteland areas and Gobi desert areas; the other is "secondary bare surface", which refers to the situation in which the original vegetation community is eliminated due to external interference such as human activities, resulting in the bare surface, such as soil erosion areas, karst desert areas, and urban built-up areas.

[0003] The state and degree of surface exposure can be used as an indicator of regional ecological fragility or environmental deterioration risk. Relatively speaking, the more obvious the surface exposure is, the less vegetation is covered in the spatial area, and the vulnerability and deterioration risk of the ecological environment are often higher; while the lower the degree of surface exposure is, the better the vegetation growth in the spatial area is, the stronger the ability to resist external climate and human interference is, the faster the recovery speed after being damaged, and the vulnerability and deterioration risk of the ecological environment are often lower. Therefore, monitoring the exposed surface fraction (ESF) within a certain time and space range has important guiding significance for assessing the regional ecological environment.

[0004] Previously, some related studies have carried out monitoring of exposed surfaces such as rocky desertification carbonate rocks, grassland degradation in alpine areas, and impermeable layers in built-up areas, but they mainly use manual visual interpretation or ordinary remote sensing calculation methods, which have the disadvantages of large data volume, complex processes, low accuracy, high storage requirements, and slow operation speed. In addition, due to the limitations of human, financial, and material resources, it is difficult to carry out large-scale and long-term ESF monitoring. This results in the evolution of the exposed surface and related information cannot be grasped in a timely and accurate manner, and cannot effectively guide the improvement of the ecological environment.

[0005] Google Earth Engine (GEE) is a cloud computing platform dedicated to processing satellite images and other earth observation data. It uses Google's most advanced cloud computing and storage capabilities to deal with various social hot issues, including deforestation, drought spread, disaster forecasting, disease prevention and control, food security, resource management, climate monitoring and environmental protection. GEE puts data storage, computing operations, integrated transmission and other processes on the cloud, and has long-term remote sensing data and related products from around the world, providing a good platform for rapid identification and monitoring of information features on the earth's surface. However, there is no research on using GEE to monitor large-scale, long-term exposed surfaces, so it is worth developing a method to monitor ESF within a specific time and space range. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform in response to the shortcomings of the background technology, so as to solve the problems of small spatial scope, short time limit, difficult data integration, serious resource consumption, complicated processing flow, insufficient accuracy and low automation level in the previous bare surface monitoring process.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform includes the following steps:

[0009] Step 1: In the GEE cloud platform, use JavaScript programming language to set the geographical boundaries where ESF monitoring is required in the form of spatial coordinates as the monitoring area range, where ESF represents the surface exposure rate; visualize the monitoring area range so that the monitoring area range can be directly seen in the GEE interface;

[0010] Step 2: In the GEE cloud platform, use the JavaScript programming language to define the cloud removal / shadow removal function of the Landsat series satellite remote sensing image data;

[0011] The remote sensing data used include Landsat-4 / 5TM satellite images and Landsat-8OLI satellite images. The cloud removal / shadow removal functions of these two types of image data are defined respectively for cloud and shadow removal preprocessing of large-scale and long-time series batch remote sensing image data.

[0012] Step 3: In the GEE cloud platform, JavaScript programming language is used to define the monitoring function for monitoring the surface exposure rate ESF from remote sensing images through the normalized surface exposure index NDBI and the pixel binary model calculation formula;

[0013] Step 4: In the GEE cloud platform, write a "for" loop statement using the JavaScript programming language and set the start and end time range of ESF monitoring; combined with the monitoring area range set in step 1, load all Landsat-4 / 5TM and Landsat-8OLI satellite remote sensing images within the start and end time range and the monitoring area from the GEE cloud platform, where only images with cloud cover less than 40% are required to be loaded; in addition, since the spectral sorting in the multispectral data of Landsat-4 / 5TM and Landsat-8OLI satellite images is inconsistent, the spectral bands of the two data are uniformly sorted, and the "blue / green / red / near infrared / short-wave infrared" bands of the two data are uniformly named "B1 / B2 / B3 / B4 / B5" for subsequent remote sensing index calculations;

[0014] Step 5: In the GEE cloud platform, the remote sensing image data obtained in step 4 and uniformly named are preprocessed according to the cloud removal / shadow removal function in step 2; the "shortwave infrared-B5" and "near infrared-B4" bands of the preprocessed remote sensing images are called to calculate the NDBI remote sensing index of each image, where NDBI represents the normalized surface exposure index value; the NDBI remote sensing index of each image is spatially superimposed year by year, and the annual average value of the NDBI index of the monitoring area is calculated;

[0015] Step 6. In the same way as step 5, in the GEE cloud platform, call the pre-processed remote sensing image "near infrared-B4" and "red light-B3" bands to calculate the annual average value of the normalized vegetation index NDVI in the monitoring area. Since a negative NDVI index indicates that the area is a water body without vegetation, the water body in the monitoring area is identified by the NDVI index average value being less than zero. Then, combined with the annual average value of the NDBI index in step 5, the NDBI index value mask of the water body is set to zero, indicating that there is no surface exposure in the water body area, so as to improve the accuracy of monitoring ESF through the NDBI index.

[0016] Step 7: In the GEE cloud platform, call the ESF monitoring function of step 3, bring the NDBI index after water masking into the function, and monitor the ESF results of the region; draw the frequency statistical histogram of the NDBI remote sensing index after water masking and the surface exposure rate for frequency distribution analysis of surface exposure conditions;

[0017] Step 8: Call the Google Drive interface of the GEE cloud platform, export the surface exposure rate monitoring result data and products within the set time and space range according to the specified spatial resolution and geographic coordinate system, and use geographic information software for mapping analysis and preservation.

[0018] Furthermore, in step 2, the cloud layer and shadow information codes of the Landsat series satellite image data on the GEE cloud platform are used to define functions for removing cloud / shadow preprocessing for Landsat-4 / 5TM data and Landsat-8OLI data using the JavaScript programming language, thereby improving the accuracy of subsequent NDBI and NDVI remote sensing index calculation results; among them, the cloud layer code of Landsat-4 / 5TM images in the GEE cloud platform is 5, and when the code is 7, it means that the probability of the pixel being a cloud layer is relatively high, and the shadow code is 3; deleting this type of pixels will obtain the image data after cloud / shadow removal; the cloud layer code of Landsat-8OLI images in the GEE platform is 5, and the shadow code is 3. Similarly, deleting this type of pixels will obtain the image data after cloud / shadow removal.

[0019] Furthermore, in step three, ESF needs to be calculated through NDBI and pixel dichotomy model, so the monitoring function for monitoring ESF through NDBI and pixel dichotomy model is defined in the GEE cloud platform using JavaScript programming language;

[0020] The ESF calculation formula is as follows:

[0021]

[0022] Where, ESF is the surface exposure rate of a specific spatial pixel, NDBI is the normalized surface exposure index value of the spatial pixel, and NDBI is max and NDBI min are the maximum and minimum values ​​of NDBI in the monitoring area, respectively.

[0023] Furthermore, when determining the maximum and minimum NDBI values ​​in the monitoring area, the extreme values ​​below 5% in the low-value area and above 95% in the high-value area of ​​NDBI are eliminated through image noise reduction processing; that is, all the values ​​below the lowest 5% of the NDBI in the monitoring area are set as minimum values, and all the values ​​of the highest 95-100% of the NDBI in the monitoring area are set as maximum values, so that the calculation is more accurate when monitoring ESF, avoiding the situation where some extreme conditions in a large scale range interfere with the monitoring results.

[0024] Furthermore, in step 4, a "for" yearly loop statement is written and the starting range of the long time series is set. All Landsat-4 / 5TM satellite and Landsat-8OLI satellite remote sensing image data within the time and space range are collected year by year through the ee.ImageCollection command of the GEE cloud platform. The data are superimposed and supplemented with each other to ensure the accuracy of subsequent index calculations and ESF monitoring.

[0025] At the same time, the inconsistent band names of Landsat-4 / 5TM and Landsat-8OLI image data are unified into a single list STD_BANDS to ensure consistency during band calculation; ['B1', 'B2', 'B3', 'B4', 'B5'] are used to represent [Blue band, Green band, Red band, NIR near infrared band, SWIR short wave infrared band] of Landsat remote sensing image multispectral data respectively.

[0026] Furthermore, in step five, the collected remote sensing images are first pre-processed by removing clouds and shadows, and then the NDBI remote sensing index of each image is calculated by the calculation formula of the uniformly named 'B5'-SWIR band and 'B4'-NIR band; the NDBI index results of all images year by year are spatially superimposed and the average value is calculated to obtain the NDBI remote sensing index calculation results of each year within the set time and space range; this operation process is to calculate the NDBI first and then stitch the images and calculate the average value. Compared with the method of stitching the remote sensing images first and then calculating the NDBI, this operation reduces the error problems caused by different weather conditions, seasonal changes, light angles, etc. between images;

[0027] The band calculation formula of NDBI is as follows:

[0028]

[0029] Wherein, NDBI is the normalized land exposure index value of a specific spatial pixel, SWIR is the shortwave infrared band value of the remote sensing image, that is, B5 after the unified band naming; NIR is the near infrared band value of the remote sensing image, that is, B4 after the unified band naming.

[0030] Furthermore, in step six, in order to avoid the inaccurate value of NDBI calculation results in the water area and interfere with ESF monitoring, the NDBI value of the water area is directly set to zero; all water bodies in the monitoring area are identified by the value of the normalized vegetation index NDVI less than zero, and then the NDBI index value of the water area is directly masked to zero, eliminating the interference of water bodies in the monitoring process and making the monitoring results more accurate;

[0031] The band calculation formula of NDVI is as follows:

[0032]

[0033] Wherein, NDVI is the normalized vegetation index value of a specific spatial pixel. When NDVI < 0, the spatial pixel area is a water body. SWIR is the near infrared band value of the remote sensing image, that is, B4 after the unified band naming. NIR is the red light band value of the remote sensing image, that is, B3 after the unified band naming.

[0034] Furthermore, in step seven, the monitoring function defined in the previous step three and the final result of NDBI are called to monitor the ESF results of each year within the set time and space range, and the frequency histogram statistics of the NDBI and ESF results are performed on the GEE cloud platform to assist in the analysis of the spatiotemporal changes of the surface exposure rate monitoring process and results.

[0035] Furthermore, in step eight, the ESF data results are automatically retrieved through the Google Drive interface, the spatial resolution and geographic coordinate system are set according to the needs, and the ESF results of each year within the time and space range are directly exported to the local computer; conventional geographic information software is used for mapping and storage to obtain the annual distribution results of the surface exposure rate within the monitoring range, and to achieve large-scale and long-term surface exposure rate monitoring based on the GEE cloud platform, which is used for the analysis of the spatiotemporal changes in the degree of surface exposure and ecological environment conditions.

[0036] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0037] 1. Previous studies have not carried out surface exposure rate monitoring in a large-scale, long-term spatial and temporal range. Existing studies have only calculated the NDBI index through ordinary remote sensing image processing, and monitored surface exposure characteristics such as desertification, grassland degradation, and impermeable layers. On the one hand, these scattered surface exposure conditions have not been integrated for unified monitoring, and on the other hand, large-scale and long-term surface exposure monitoring has not been carried out through the most cutting-edge remote sensing cloud platform means. Therefore, the present invention, with the help of the GEE cloud platform, directly performs preprocessing and basic calculations of remote sensing image data in the cloud, thereby improving the efficiency of surface exposure rate monitoring, avoiding the instability of manual data processing operation procedures, and ensuring the high accuracy of monitoring results.

[0038] 2. The current Landsat series of remote sensing satellites perform strip scanning of the earth's surface information, and each scan width does not exceed 200km. Therefore, the remote sensing images collected in a large-scale area are not taken on the same day, resulting in color difference between images. This method uses the method of first calculating the NDBI remote sensing index of all remote sensing images, and then obtaining the NDBI average value through overlay analysis, which avoids this color difference problem as much as possible. It is more stable and accurate than the current method of first overlaying and splicing remote sensing images and then calculating the remote sensing index.

[0039] 3. Since water bodies interfere with the NDBI calculation and surface exposure rate monitoring process, and water bodies themselves should not be characterized as bare surfaces, this method identifies and eliminates the interference of water bodies on NDBI through the negative value interval of the vegetation index NDVI, and directly sets the surface exposure rate of the water area to zero, which is conducive to improving the accuracy of the final ESF monitoring results.

[0040] 4. Based on the remote sensing big data of the GEE cloud platform, a large-scale and long-time series surface exposure rate monitoring method suitable for land remote sensing was constructed through satellite image band information calculation and denoising. The monitoring results can be synchronized to the local area for mapping and analysis, providing data support for regional long-term ecological environment assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a system flow chart of the method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform;

[0042] Figure 2 is a frequency statistical histogram of the NDBI calculation results for 2021 in the embodiment;

[0043] Figure 3 is a frequency statistical histogram of the surface exposure rate monitoring results in 2021 in the embodiment;

[0044] Figure 4 is the surface exposure rate monitoring result in 1989 in the embodiment (unit: %);

[0045] Figure 5 It is the surface exposure rate monitoring result in 2021 in the embodiment (unit: %). DETAILED DESCRIPTION

[0046] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings:

[0047] like Figure 1 As shown in FIG. 1 , the method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform includes the following steps 1 to 8:

[0048] Step 1: Set the monitoring boundary range and draw it visually on the GEE cloud platform.

[0049] In this embodiment, the present invention is based on the GEE cloud platform, and sets the monitoring range to 89.58037741946824, 39.43134583007215, 117.5373675577281, 19.46312680663627 (i.e., the upper and lower limits of the longitude and latitude of the boundary). And the boundary of the range is visualized on the GEE cloud platform, and the color is set to white. The specific range setting embodiment code is as follows:

[0050]

[0051] By running the above JavaScript programming language in the GEE cloud platform, you can get the spatial monitoring range "JCRegion" area of ​​this embodiment, and display it in light gray in the GEE cloud platform.

[0052] Step 2: Define the cloud / shadow removal function for Landsat series remote sensing image data.

[0053] In this embodiment, cloud removal operations are performed on Landsat-4 / 5TM data and Landsat-8OLI data.

[0054] For Landsat-4 / 5TM data, in the code, when 'pixel_qa' = (5), it means that the pixel is in the cloud position, and when 'pixel_qa' = (7), it means that the confidence of the pixel being in the cloud layer is high, so this type of pixel is eliminated; in addition, when 'pixel_qa' = (3), it means that the pixel is in the shadow and is a bad pixel, so this type of pixel is also eliminated. By using the function function of the JavaScript programming language, the specific function setting embodiment code is as follows:

[0055]

[0056] For Landsat-8OLI data, when 'pixel_qa' = (3) and 'pixel_qa' = (5), it means that the pixel is a cloud shadow and a cloud, respectively, so the pixel values ​​of this type are also set to zero. The specific function setting embodiment code is as follows:

[0057]

[0058] Step three: define the function of monitoring the surface exposure rate through the NDBI index and the pixel binary model.

[0059] In this embodiment, a function for calculating the surface exposure rate by using NDBI and pixel binary model is defined, with specific reference to the aforementioned ESF calculation formula. In the function, it is required to normalize the values ​​in the range of 5% to 95% of the NDBI index, remove the extreme values, and then use them to monitor the surface exposure rate. The definition of this function uses JavaScript programming languages ​​such as ee.Reducer and ee.Number. The code of the embodiment for defining the function is as follows:

[0060]

[0061] Step 4: Set the start and end time nodes for ESF monitoring. According to the time limit and the spatial scope of step 1, load the Landsat series satellite images from the GEE cloud platform, and pre-process the remote sensing images with related functions and codes.

[0062] In this embodiment, 1989 and 2021 are set as the starting and ending points of the monitoring time, and then the satellite remote sensing images are collected and preprocessed year by year from 1989 to 2021 in combination with the spatial range in step 1 (only remote sensing images with cloud cover below 40% are collected). This step is mainly implemented by the "for" structure loop statement of the JavaScript programming language. The specific implementation code of a series of operations such as time setting, image loading, cloud cover filtering, cloud removal processing, and band naming is as follows:

[0063]

[0064]

[0065] So far, this embodiment has acquired all Landsat-4 / 5TM and Landsat-8OLI remote sensing image data within the monitoring area from 1989 to 2021 with cloud cover below 40%, completed cloud removal / shadow processing, and unified the band naming as required.

[0066] Step 5: Perform band calculations on the acquired images through the GEE cloud platform to obtain the average value of the NDBI remote sensing index in the monitoring area and period.

[0067] In this embodiment, the NDBI remote sensing index is directly generated by renaming the short-wave infrared SWIR band 'B5' and the near-infrared NIR band 'B4' using the band calculation formula ('B5'–'B4') / ('B5'+'B4'). The specific embodiment code is as follows (the code needs to be used in the for loop statement in step 4):

[0068] var NDBI_chushi=image.normalizedDifference(["B5","B4"]).rename("NDBI_chushi");

[0069] return image.addBands(NDBI_chushi);

[0070] .select("NDBI_chushi");

[0071] Then, the NDBI calculated from all images is spatially superimposed to obtain the average NDBI value of the monitoring area year by year. The specific implementation code is as follows:

[0072]

[0073]

[0074] So far, this embodiment has obtained the annual average NDBI value within the monitoring area from 1989 to 2021.

[0075] Step 6: Use the same method as step 5 to calculate the average NDVI value in the GEE cloud platform, and identify the water body range in the monitoring area based on the situation of NDVI < 0; then set the NDBI index value in the water body area to zero.

[0076] In this embodiment, the NDVI remote sensing index is directly generated by renaming the near infrared NIR band 'B4' and the red band 'B3' using the band calculation formula ('B4'–'B3') / ('B4'+'B3'), and all NDVI values ​​< 0 are identified as water bodies. The specific embodiment code is as follows:

[0077]

[0078] Then the NDBI average value mask within the water body range is set to zero. The specific implementation code is as follows:

[0079] Map.addLayer(NDBI.updateMask(NDVI.gt(0)),ndbi_vis,'NDBI'+year);

[0080] var mask_NDBI=NDBI.updateMask(NDVI.gt(0));

[0081] So far, this embodiment obtains the water body mask NDBI results of the monitoring area year by year, that is, the upper mask_NDBI.

[0082] Step 7: Use the GEE cloud platform to monitor the surface exposure rate using NDBI with a water body range mask of zero, and draw a statistical chart of NDBI and surface exposure rate.

[0083] In this embodiment, the NDBI data after masking in step 6 is used, and the pixel binary model function defined in step 3 is introduced to calculate and generate the surface exposure rate result. In this step, the monitoring result is only stored and displayed in the GEE cloud platform.

[0084] Then, the Chart.image.histogram command in the GEE cloud platform is called to perform frequency histogram statistics on the spatial extraction results of NDBI and surface exposure rate to obtain the frequency distribution for the evaluation and analysis of the relevant ecological environment status. The specific implementation code is as follows:

[0085]

[0086]

[0087] Now we can analyze the statistical chart. Figure 2 The frequency statistical histogram of NDBI in Example 2021 is shown. The NDBI values ​​are mainly between -0.45 and 0.25, among which the number of pixels near the NDBI value of 0.05 is the largest (greater than 800,000), followed by the number of pixels near the NDBI value of -0.25 (about 784,475); correspondingly, Figure 3 The ESF frequency statistical histogram of the embodiment 2021 calculated by NDBI is shown. The ESF monitoring results are between 0 and 1, among which the number of pixels with ESF of 0 is the largest (greater than 500,000), indicating that the surface exposure rate in most areas within the monitoring range is zero, that is, there is no surface exposure in these areas, either dense vegetation coverage or water areas, and the ecological environment is in good condition. Secondly, the number of pixels with ESF of 0.9 is also large (about 403,453), and the surface exposure rate in the areas where these pixels are distributed is obvious, and the vegetation coverage is less, especially when ESF is 1, it means that the surface is completely exposed and the vegetation coverage is missing, the ecological environment is poor, and corresponding ecological restoration and improvement measures are urgently needed.

[0088] Step 8. Call the Google Drive interface of the GEE cloud platform to export the annual surface exposure rate monitoring results data and products from 1989 to 2021 to the local computer, and perform mapping analysis and save them.

[0089] In this embodiment, the generated surface exposure rate is exported to the local computer through the call of the Export.image.toDrive command. The specific embodiment code is as follows:

[0090]

[0091] Then, through ArcGIS and other geospatial analysis software, the surface exposure rate can be mapped and further analyzed. Figure 4 and 5The results of spatial mapping of land surface exposure rate in 1989 and 2021 are shown (the legend in the lower left corner indicates the monitoring year and the expression of land surface exposure rate). It can be seen that the land surface exposure rate in the north, west, northwest, southwest and other regions of the monitoring range is relatively high. These regions are the key distribution areas of China's Qinghai-Tibet Plateau, grassland degradation, Gobi desertification, karst rocky desertification, etc., and the ecological environment is poor, which is in line with geographical laws and actual conditions; while the land surface exposure rate in Chongqing, Hubei, Hunan, Jiangxi, Guangdong, Guangxi and other regions is relatively low, indicating that these provinces have good vegetation coverage and superior ecological environment conditions. From the perspective of dynamic evolution, comparing the monitoring results of land surface exposure rate in 1989 and 2021, the most obvious changes are in the eastern part of Yunnan Province, the western part of Guizhou Province, the central part of Guangxi Zhuang Autonomous Region, the eastern part of Guangdong Province, the central part of Hunan Province, and the western part of Hubei Province. The land surface exposure rate in these regions was relatively high in 1989, while the land surface exposure rate in 2021 was relatively low, indicating that the ecological environment has been significantly improved during this period. Compared with the actual situation, it is consistent with the actual situation, which shows that the large-scale and long-time series surface exposure rate monitoring results of the present invention are accurate and reasonable.

[0092] At this point, the rapid and automated annual monitoring and analysis of the surface exposure rate in the monitoring area from 1989 to 2021 in this embodiment has been completed.

[0093] The specific explanations of the Chinese and English abbreviations used in this embodiment are as follows:

[0094]

[0095] The key points and protection points of the present invention are to call the remote sensing big data of the GEE cloud platform through code, perform operations such as de-clouding, noise reduction, and calculation of the band information of satellite images, and construct a large-scale and long-time series surface exposure rate monitoring method suitable for land remote sensing, providing data support for regional long-term ecological environment assessment.

[0096] The advantages or beneficial effects of the present invention are as follows:

[0097] 1. Previous studies have not carried out surface exposure rate monitoring in a large-scale, long-term spatial and temporal range. Existing studies have only calculated the NDBI index through ordinary remote sensing image processing, and monitored surface exposure characteristics such as desertification, grassland degradation, and impermeable layers. On the one hand, these scattered surface exposure conditions have not been integrated for unified monitoring, and on the other hand, large-scale and long-term surface exposure monitoring has not been carried out through the most cutting-edge remote sensing cloud platform means. Therefore, the present invention, with the help of the GEE cloud platform, directly performs preprocessing and basic calculations of remote sensing image data in the cloud, thereby improving the efficiency of surface exposure rate monitoring, avoiding the instability of manual data processing operation procedures, and ensuring the high accuracy of monitoring results.

[0098] 2. The current Landsat series of remote sensing satellites perform strip scanning of the earth's surface information, and each scan width does not exceed 200km. Therefore, the remote sensing images collected in a large-scale area are not taken on the same day, resulting in color difference between images. This method uses the method of first calculating the NDBI remote sensing index of all remote sensing images, and then obtaining the NDBI average value through overlay analysis, which avoids this color difference problem as much as possible. It is more stable and accurate than the current method of first overlaying and splicing remote sensing images and then calculating the remote sensing index.

[0099] 3. Since water bodies interfere with the NDBI calculation and surface bare rate monitoring process, and water bodies themselves should not be characterized as bare surfaces, this method uses vegetation index to identify and eliminate the interference of water bodies on NDBI, and directly sets the surface bare rate of the water area to zero, which is conducive to improving the accuracy of the final ESF monitoring results.

[0100] 4. Based on the remote sensing big data of the GEE cloud platform, a large-scale and long-time series surface exposure rate monitoring method suitable for land remote sensing was constructed through satellite image band information calculation and denoising. The monitoring results can be synchronized to the local area for mapping and analysis, providing data support for regional long-term ecological environment assessment.

[0101] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0102] The above embodiments are only for explaining the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the present invention. The above embodiments of the present invention are described in detail, but the present invention is not limited to the above embodiments. Various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention.

Claims

1. A method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform, characterized by: The steps include: Step 1: In the GEE cloud platform, use JavaScript programming language to set the geographical boundaries where ESF monitoring is required in the form of spatial coordinates as the monitoring area range, where ESF represents the surface exposure rate; visualize the monitoring area range, so that the monitoring area range can be directly seen in the interface of the GEE cloud platform; Step 2: In the GEE cloud platform, use the JavaScript programming language to define the cloud removal / shadow removal function of the Landsat series satellite remote sensing image data; The remote sensing data used include Landsat-4 / 5 TM satellite images and Landsat-8 OLI satellite images. The cloud removal / shadow removal functions of these two types of image data are defined respectively for cloud and shadow removal preprocessing of large-scale and long-time series batch remote sensing image data. Step 3: In the GEE cloud platform, JavaScript programming language is used to define the monitoring function for monitoring the surface exposure rate ESF from remote sensing images through the normalized surface exposure index NDBI and the pixel binary model calculation formula; ESF is calculated by NDBI and pixel dichotomy model, so the function of monitoring ESF by NDBI and pixel dichotomy model is defined in GEE cloud platform using JavaScript programming language; The ESF calculation formula is as follows: ， In the formula, ESF is the surface exposure rate of a specific spatial pixel, NDBI is the normalized surface exposure index value of the spatial pixel, NDBI max and NDBI min Within the monitoring area NDBI The maximum and minimum values ​​of ; Step 4: In the GEE cloud platform, write a "for" loop statement using the JavaScript programming language and set the start and end time range of ESF monitoring; combined with the monitoring area range set in step 1, load all Landsat-4 / 5 TM and Landsat-8 OLI satellite remote sensing images within the start and end time range and the monitoring area from the GEE cloud platform, and only load images with cloud cover less than 40%; in addition, since the spectral sorting in the multispectral data of Landsat-4 / 5 TM and Landsat-8 OLI satellite images is inconsistent, the spectral bands of the two data are uniformly sorted, and the "blue / green / red / near infrared / shortwave infrared" bands of the two data are uniformly named "B1 / B2 / B3 / B4 / B5" for subsequent remote sensing index calculations; Step 5: In the GEE cloud platform, the remote sensing image data obtained in step 4 and uniformly named are preprocessed according to the cloud removal / shadow removal function in step 2; the "shortwave infrared-B5" and "near infrared-B4" bands of the preprocessed remote sensing images are called to calculate the NDBI remote sensing index of each image, where NDBI represents the normalized surface exposure index value; the NDBI remote sensing index of each image is spatially superimposed year by year, and the annual average value of the NDBI index of the monitoring area is calculated; Step 6. In the same way as step 5, in the GEE cloud platform, call the pre-processed remote sensing image "near infrared-B4" and "red light-B3" bands to calculate the annual average value of the normalized vegetation index NDVI in the monitoring area. Since a negative NDVI index indicates that the area is a water body without vegetation, the water body in the monitoring area is identified by the NDVI index average value being less than zero. Then, combined with the annual average value of the NDBI index in step 5, the NDBI index value mask of the water body is set to zero, indicating that there is no surface exposure in the water body area, so as to improve the accuracy of monitoring ESF through the NDBI index. Step 7: In the GEE cloud platform, call the ESF monitoring function of step 3, bring the NDBI index after water masking into the function, and monitor the ESF results of the region; draw the frequency statistical histogram of the NDBI remote sensing index after water masking and the surface exposure rate for frequency distribution analysis of surface exposure conditions; Step 8: Call the Google Drive interface of the GEE cloud platform, export the surface exposure rate monitoring result data and products within the set time and space range according to the specified spatial resolution and geographic coordinate system, and use geographic information software for mapping analysis and preservation.

2. The method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform according to claim 1 is characterized by: Step 2: Through the cloud layer and shadow information codes of the Landsat series satellite image data on the GEE cloud platform, the JavaScript programming language is used to define the functions for removing cloud / shadow preprocessing for Landsat-4 / 5 TM data and Landsat-8 OLI data, thereby improving the accuracy of the subsequent NDBI and NDVI remote sensing index calculation results; among them, the cloud layer code of Landsat-4 / 5 TM images in the GEE cloud platform is 5, and when the code is 7, it means that the probability of the pixel being a cloud layer is high, and the shadow code is 3; deleting this type of pixels will obtain the image data after cloud / shadow removal; the cloud layer code of Landsat-8 OLI images in the GEE cloud platform is 5, and the shadow code is 3. Similarly, deleting this type of pixels will obtain the image data after cloud / shadow removal.

3. According to claim 2, a method for monitoring large-scale and long-time series surface exposure rate based on GEE cloud platform is characterized in that: When determining the maximum and minimum NDBI values ​​in the monitoring area, the image noise reduction process is used to eliminate extreme values ​​below 5% in the low-value area and above 95% in the high-value area of ​​NDBI; that is, all values ​​below the lowest 5% of the NDBI in the monitoring area are set as minimum values, and all values ​​between 95% and 100% of the highest NDBI in the monitoring area are set as maximum values, so that the calculation is more accurate when monitoring ESF and avoids the situation where some extreme conditions in a large scale range interfere with the monitoring results.

4. The method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform according to claim 1 is characterized by: Step 4: Write a "for" year-by-year loop statement and set the starting range of the long time series. Use the ee.ImageCollection command of the GEE cloud platform to collect all Landsat-4 / 5 TM satellite and Landsat-8 OLI satellite remote sensing image data within the temporal and spatial range. The data are superimposed and supplemented to ensure the accuracy of subsequent index calculations and ESF monitoring. At the same time, the inconsistent band names of Landsat-4 / 5 TM and Landsat-8 OLI image data are unified into a single list STD_BANDS to ensure consistency in the band operation process; ['B1', 'B2', 'B3', 'B4', 'B5'] are used to represent [Blue band, Green band, Red band, NIR near infrared band, SWIR short wave infrared band] of Landsat remote sensing image multispectral data respectively.

5. According to claim 1, a method for monitoring large-scale and long-time series surface exposure rate based on GEE cloud platform is characterized by: Step 5: First, the collected remote sensing images were preprocessed to remove clouds and shadows, and then the NDBI remote sensing index of each image was calculated using the calculation formula of the uniformly named 'B5'-SWIR band and 'B4'-NIR band. The NDBI index results of all images year by year were spatially superimposed and the average value was calculated to obtain the NDBI remote sensing index calculation results of each year within the set time and space range. This operation process is to calculate the NDBI first and then stitch the images and calculate the average value. Compared with the method of stitching remote sensing images first and then calculating the NDBI, this operation reduces the error problems caused by weather conditions, seasonal changes, and different light angles between images. The band calculation formula of NDBI is as follows: ， In the formula, NDBI is the normalized land exposure index value of a specific spatial pixel, SWIR It is the short-wave infrared band value of the remote sensing image, that is, B5 after the unified band naming; NIR It is the near-infrared band value of the remote sensing image, that is, B4 after the unified band naming.

6. The method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform according to claim 1 is characterized by: Step 6: To avoid inaccurate NDBI calculation results in water areas and interference with ESF monitoring, the NDBI value of the water area is directly set to zero; all water bodies in the monitoring area are identified by the value of the normalized vegetation index NDVI less than zero, and then the NDBI index value of the water area is directly masked to zero, eliminating the interference of water bodies in the monitoring process and making the monitoring results more accurate; The band calculation formula of NDVI is as follows: ， In the formula, NDVI is the normalized vegetation index value of a specific spatial pixel. When NDVI < 0, the spatial pixel area is a water body. SWIR is the near infrared band value of the remote sensing image, that is, B4 after the unified band naming, NIR It is the red light band value of the remote sensing image, that is, B3 after the unified band naming.

7. The method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform according to claim 1 is characterized by: Step 7: Call the monitoring function defined in the previous step 3 and the final result of NDBI to monitor the ESF results each year within the set time and space range, and perform frequency histogram statistics on the NDBI and ESF results in the GEE cloud platform to assist in the analysis of the spatiotemporal changes in the surface exposure rate monitoring process and results.

8. The method for monitoring large-scale and long-time series surface exposure rate based on the GEE cloud platform according to claim 1 is characterized by: Step 8: Automatically retrieve ESF data results through the Google Drive interface, set the spatial resolution and geographic coordinate system according to needs, and directly export the ESF results of each year within the time and space range to the local computer; use conventional geographic information software for mapping and storage to obtain the annual distribution results of the surface exposure rate within the monitoring range, and achieve large-scale and long-term surface exposure rate monitoring based on the GEE cloud platform, which is used for the analysis of the spatiotemporal changes in the degree of surface exposure and ecological environment conditions.

Images