Snow Recognition Method, Device, Equipment and Storage Medium Based on Freeman Decomposition

Abstract

An embodiment of the present invention provides a snow cover recognition method, device, equipment and storage medium based on Freeman decomposition, which relates to the technical field of snow cover recognition. Among them, this snow cover recognition method includes steps S1 to S5. S1. Obtain the Radarsat-2 full-polarization image data of the target area. S2. According to the Radarsat-2 full-polarization image data, perform Freeman polarization decomposition to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data. S3. Construct a snow cover index for the Radarsat-2 full-polarization image data according to the dihedral angle scattering characteristic and the volume scattering characteristic. S4. Statistically analyze the snow cover sample histogram according to the snow cover index, and obtain the recognition threshold according to the snow cover sample histogram. S5. According to the recognition threshold, judge the types of each pixel in the snow cover index to obtain the snow cover spatial distribution map of the target area. Among them, the types include snow cover and non-snow cover. The snow cover recognition method can quickly and accurately identify the snow cover area.

Description

Technical Field

[0001] The present invention relates to the technical field of snow cover recognition, and in particular, to a snow cover recognition method, device, equipment and storage medium based on Freeman decomposition. Background Art

[0002] Snow cover is the most widely distributed and active element in the cryosphere and plays an important role in the hydrological cycle, snow disaster prediction, etc. Using optical remote sensing data for snow cover research is susceptible to the influence of cloudy and rainy weather. Synthetic Aperture Radar (SAR) can penetrate clouds and has the characteristics of all-weather and all-time imaging, and has great potential in the application of mountain snow cover recognition and classification.

[0003] Full-polarization SAR data can reflect the polarization scattering characteristics of snow cover and distinguish the scattering mechanisms of ground objects, providing rich polarization information for snow cover recognition and helping to improve the accuracy of snow cover recognition. Extracting effective polarization features from SAR images is of great significance for the research and analysis of snow cover characteristics.

[0004] However, there is still much room for improvement in the information extraction of polarization features obtained by single polarization decomposition, and how to quickly obtain the snow cover recognition range using less scattering information still needs further exploration.

[0005] In view of this, the applicant has put forward this application after studying the existing technologies. Summary of the Invention

[0006] The present invention provides a snow cover recognition method, device, equipment and storage medium based on Freeman decomposition to improve at least one of the above technical problems.

[0007] First Aspect,

[0008] An embodiment of the present invention provides a snow cover recognition method based on Freeman decomposition, which includes steps S1 to S5.

[0009] S1. Obtain the Radarsat-2 full-polarization image data of the target area.

[0010] S2. Perform Freeman polarization decomposition on the Radarsat-2 full-polarization image data to obtain the polarization features of each pixel in the Radarsat-2 full-polarization image data. Among them, the polarization features include the dihedral angle scattering feature Freeman_Dbl and the volume scattering feature Freeman_Vol.

[0011] S3. Construct the snow cover index FscI of the Radarsat-2 full-polarization image data according to the dihedral angle scattering feature and the volume scattering feature_vol_Dbl Among them,

[0012] S4. Statistically analyze the snow cover sample histogram based on the snow cover index, and obtain the recognition threshold according to the snow cover sample histogram.

[0013] S5. Determine the type of each pixel in the snow cover index according to the recognition threshold to obtain the snow cover spatial distribution map of the target area. Among them, the types include snow cover and non - snow cover.

[0014] In the second aspect,

[0015] An embodiment of the present invention provides a snow cover recognition device based on Freeman decomposition, which includes:

[0016] An initial data acquisition module, configured to acquire Radarsat - 2 full - polarization image data of the target area.

[0017] A decomposition module, configured to perform Freeman polarization decomposition on the Radarsat - 2 full - polarization image data to obtain the polarization characteristics of each pixel in the Radarsat - 2 full - polarization image data. Among them, the polarization characteristics include the dihedral angle scattering characteristic Freeman_Dbl and the volume scattering characteristic Freeman_Vol.

[0018] An index construction module, configured to construct the snow cover index FSCI of the Radarsat - 2 full - polarization image data according to the dihedral angle scattering characteristic and the volume scattering characteristic _Vol_Dbl Among them,

[0019] A threshold acquisition module, configured to statistically analyze the snow cover sample histogram based on the snow cover index, and obtain the recognition threshold according to the snow cover sample histogram.

[0020] A snow cover recognition module, configured to determine the type of each pixel in the snow cover index according to the recognition threshold to obtain the snow cover spatial distribution map of the target area. Among them, the types include snow cover and non - snow cover.

[0021] In the third aspect,

[0022] An embodiment of the present invention provides a snow cover recognition device based on Freeman decomposition, which includes a processor, a memory, and a computer program stored in the memory. The computer program can be executed by the processor to implement the snow cover recognition method based on Freeman decomposition as described in any paragraph of the first aspect.

[0023] In the fourth aspect,

[0024] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the snow cover recognition method based on Freeman decomposition described in any paragraph of the first aspect.

[0025] By adopting the above technical solution, the present invention can achieve the following technical effects:

[0026] The snow cover recognition method according to the embodiment of the present invention can quickly and accurately identify the snow-covered area based on the full-polarization synthetic aperture radar image.

[0027] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following specifically enumerates preferred embodiments and, in conjunction with the accompanying drawings, makes a detailed description as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and thus should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0029] Figure 1 is a schematic flowchart of the snow cover recognition method provided by the first embodiment of the present invention.

[0030] Figure 2 is the result after Freeman polarization decomposition of the Radarsat-2 full-polarization data in the study area.

[0031] Figure 3 is Figure 2 the Freeman three-component profile during the snow-free period in area A in

[0032] Figure 4 is Figure 2 the Freeman three-component profile during the dry snow period in area A in

[0033] Figure 5 is Figure 2 the Freeman three-component profile during the snow-free period in area B in

[0034] Figure 6 is Figure 2 the Freeman three-component profile during the dry snow period in area B in

[0035] Figure 7 is the probability distribution density of the scattering characteristics in the snow-covered area.

[0036] Figure 8 is the probability distribution density of the scattering characteristics in the snow-free area.

[0037] Figure 9 is a schematic structural diagram of the snow cover recognition device provided by the second embodiment of the present invention. Detailed implementation manners

[0038] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0039] For a better understanding of the technical solutions of the present invention, the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0040] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms of "a", "the" and "said" used in the embodiments of the present invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0041] It should be understood that the term " / and" used herein is only a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after.

[0042] Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "if determined" or "if detected (stated condition or event)" may be interpreted as "when determined" or "in response to determining" or "when detected (stated condition or event)" or "in response to detecting (stated condition or event)".

[0043] The "first / second" mentioned in the embodiments is only to distinguish similar objects, and does not represent a specific order for the objects. It can be understood that the "first / second" can be interchanged in a specific order or sequence when allowed. It should be understood that the objects distinguished by "first / second" can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0044] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners:

[0045] Embodiment 1:

[0046] Please refer to Figures 1 to 8 , the first embodiment of the present invention provides a snow cover recognition method based on Freeman decomposition, which can be executed by a snow cover recognition device based on Freeman decomposition (hereinafter referred to as: snow cover recognition device). In particular, it is executed by one or more processors in the snow cover recognition device to implement steps S1 to S5.

[0047] S1. Obtain the Radarsat-2 full-polarization image data of the target area.

[0048] It can be understood that the snow cover recognition device can be an electronic device with computing performance such as a portable notebook computer, a desktop computer, a server, a smart phone or a tablet computer.

[0049] Preferably, step S1 is specifically:

[0050] Obtain the Radarsat-2 full-polarization image data of the target area. Among them, the Radarsat-2 full-polarization image data includes four polarization modes: HH, HV, VV, and VH.

[0051] Specifically, in order to verify the effectiveness of the embodiments of the present invention, the middle northern foot of the Tianshan Mountains in Xinjiang, the Manas River Basin is used as the study area and the research object for verification.

[0052] The longitude and latitude range of the northern foot of the Manas River Basin is 43.79°N to 44.09°N, 85.74°E to 86.13°E. The surface cover is mainly grassland. The climate belongs to the mid-temperate continental arid climate. The overall precipitation is relatively small and is greatly affected by the water vapor source, terrain and elevation. According to the natural altitude zone, the study area can be divided into three zones. More than 50% of the area has an altitude of 0-1300m, which is the transition zone from grassland to semi-shrub; about 25% of the area in the south has an altitude of 1300-2000m, which is the meadow steppe zone; a small part of the southernmost area has an altitude of 2000-2700m, which is the spruce forest zone. During the period from December 12, 2013 to December 16, 2013, ground snow cover observations synchronized with Radarast-2 satellite data were carried out in the study area for 5 days to obtain information such as snow depth, altitude, slope, air temperature, underlying surface, etc. at the observation points in the study area. According to the field observations, the snow depth range in the study area is between 0 and 20 cm.

[0053] The data sources used for verification are: two scenes of Radarsat-2 full-polarization data on October 2, 2013 and December 13, 2013, including four polarization modes of HH, HV, VV, and VH, which are single look complex (SLC) C-band products. The range resolution and azimuth resolution are 4.733 m and 4.799 m respectively, and the central incidence angle is 43.45°. Combining field observations and weather conditions, the corresponding snow cover states are snow-free and dry snow (Table 1).

[0054] Table 1 Snow cover state of Radarsat-2 data

[0055] Acquisition time Snow cover status Reference optical image and time 2013 / 10 / 02 Snow-free Landsat-8 OLI: 2013 / 10 / 06 2013 / 12 / 13 Dry snow GF-1 WFV: 2013 / 12 / 14

[0056] Meanwhile, Landsat-8 OLI data on October 6, 2013 and GF-1 WFV data on December 14, 2013 are selected for auxiliary reference. Querying the historical weather records (https: / / lishi.tianqi.com), it can be known that the temperatures from December 13 to 14, 2013 were all below 0°C and there was no rainfall record, indicating the consistency of the SAR data and optical data in the surface cover type during the dry snow period.

[0057] S2. Perform Freeman polarization decomposition on the Radarsat-2 full-polarization image data to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data. Among them, the polarization characteristics include the dihedral angle scattering characteristic Freeman_Dbl and the volume scattering characteristic Freeman_Vol.

[0058] Based on the above embodiments, in an optional embodiment of the present invention, step S2 includes steps S21 to S23.

[0059] S21. Preprocess the Radarsat-2 full-polarization image data. Among them, the preprocessing includes multi-look processing, polarization filtering, and geocoding using DEM data. Preferably, the polarization filtering is Refined Lee filtering. The DEM data is STRM4 DEM data.

[0060] Specifically, to ensure similarity to the spatial resolution of optical remote sensing data (20 m), 3×3 multi-look processing is applied to the range and azimuth directions of Radarsat-2 data respectively. Among them, the specific specifications of multi-look processing are determined by the data resolution. For fully polarized Radarsat-2 data, speckle filtering not only needs to consider the filtering of 4 channels (HH, HV, VH, VV), but also the correlation between each channel. Therefore, the Refined Lee filtering algorithm is selected for speckle suppression, and the filtering window is set to 5×5. Finally, the corresponding DEM data of the image is selected to perform geocoding on the image. In this paper, the 90-meter resolution data STRM4 DEM provided by NASA is used to perform geocoding operations on the image area, and the image is converted from slant range geometry to map projection coordinates.

[0061] S22. For the preprocessed fully polarized Radarsat-2 image data, perform polarization matrix processing to obtain the polarization scattering matrix S, and convert the polarization scattering matrix S into the polarization covariance matrix C.

[0062] Specifically, the polarization scattering matrix S can only describe the so-called coherent or pure scatterers. For distributed scatterers, to reduce the influence of speckle noise, the polarization scattering matrix S is converted into the polarization covariance matrix C3.

[0063] In the scattering process, a common method for quantitatively describing the polarization effect of SAR targets is to use the polarization scattering matrix, which contains all the polarization information of the target and is defined as follows:

[0064]

[0065] In the formula, S represents the scattering characteristics of a single pixel, S hh 、S vv represent the co-polarization terms, and S hv 、S vh represent the cross-polarization terms.

[0066] The polarization covariance matrix, like the polarization scattering matrix, contains all the target polarization information obtained by radar measurement. For a reciprocal medium S hv =S vh , the polarization scattering matrix vector can be expressed as:

[0067]

[0068] Taking a 3×3 look matrix as an example, the covariance matrix C is defined as follows:

[0069]

[0070] In the formula, * represents the conjugate matrix, T represents the transpose of the matrix, |·| represents the modulus of the matrix, and 〈·〉 represents the spatial statistical average of the random scattering medium under isotropy.

[0071] S23. Based on the polarization covariance matrix C, perform Freeman polarization decomposition to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data.

[0072] Specifically, the polarization decomposition can be realized based on the scattering matrix, the coherence matrix or the covariance matrix. The full-polarization SAR image usually uses 9 independent parameters to represent the received data, thus forming the covariance matrix or the coherence matrix. The two are equivalent and are both non-negative definite Hermitian matrices.

[0073] The Freeman decomposition takes the polarization covariance matrix as the decomposition object and divides the ground objects into three types of scattering mechanisms: volume scattering, surface scattering and dihedral angle scattering. Compared with other decomposition methods, the Freeman decomposition is more in line with the scattering mechanism of the ground objects and describes the scattering characteristics of the ground objects more fully.

[0074] The covariance matrix components of the three scattering modes are as follows:

[0075]

[0076]

[0077]

[0078] In the formula, C 1 is the covariance matrix component corresponding to the surface scattering, C 2 is the covariance matrix component corresponding to the dihedral angle scattering, C 3 is the covariance matrix component corresponding to the volume scattering; f s , f d , f v represent the contribution values of each component; α and β represent the ratios of the horizontally polarized wave to the vertically polarized wave in the polarization scattering matrix components under various scattering modes. The total polarization covariance matrix is expressed as:

[0079] C = C 1 + C 2 + C 3 (6)

[0080] From the above formulas (1) to (6), the values of the unknowns can be solved and the powers of the three scattering components can be obtained:

[0081] P S = f s (1 + |β| 2 ) (7)

[0082] P d = f d (1 + |α| 2 ) (8)

[0083] P d = f v (9) In the formula, Ps, Pd, and Pv represent the scattering powers of the surface scattering, dihedral angle scattering, and volume scattering components, respectively.

[0084] S3. According to the dihedral angle scattering characteristics and volume scattering characteristics, construct the snow cover index FSCI of Radarsat-2 full-polarization image data _Vol_Dbl . Among them,

[0085] Specifically, for the snow cover index FSCI _Vol_Dbl Using the characteristic components after Freeman polarization decomposition, a snow cover index is constructed. Through non-linear transformation, the low-value part of the index is enhanced, and the high-value part is suppressed, so that the index value generally reaches a saturated state, thereby reducing the sensitivity of the high-coverage snow area, and thus identifying snow cover.

[0086] Since both the dihedral angle scattering and volume scattering are negative values, the calculated value of FSCI _Vol_Dbl is negative, between -1 and 0, avoiding the influence on data analysis caused by too large or too small data. In addition, FSCI _Vol_Dbl can also eliminate the influence of the limitation of snow cover identification in some mountainous areas being easily affected by cloud cover. Since the characteristic components obtained by Freeman polarization decomposition are the scattering of the volume and surface, which are extremely susceptible to topographic factors, the calculated FSCI _Vol_Dbl close to -1 generally indicates a mountainous area with large terrain undulations and complex terrain, resulting in abnormal values less than -1. FSCI _Vol_Dbl close to 0 generally indicates that the area is covered by snow.

[0087] It should be noted that as Figure 2 shown, since the snow depth range in the study area is between 0 and 20 cm, which is relatively shallow, the electromagnetic wave of the C band can refract on the snow surface and penetrate dry snow. When there is snow cover, under dry snow conditions (as shown in b in Figure 2 ), the main scattering is the surface scattering at the snow-ground interface, and the volume scattering of the snow layer is less than the overall surface scattering. Combining with optical remote sensing data, the snow-covered area appears blue-violet. The snow-free area mainly appears in high-coverage grasslands and forests, and the main scattering is the volume scattering of vegetation, and the image is green.

[0088] In the Freeman decomposition results during two snow-accumulation periods, referring to the optical remote sensing images of the same period, a cross-section was taken in the study area, and the variation laws of surface scattering, dihedral-angle scattering, and volume scattering components were analyzed through the cross-section. To display as much information as possible on a single cross-section line, study area A, which included both snow-covered and non-snow-covered areas during the dry snow period and had an obvious transition section, was selected (as shown in a of Figure 2 ). Meanwhile, an area B with a large human influence (as shown in a of Figure 2 ) was selected. This area was relatively flat cultivated land.

[0089] As Figures 3 to 6 shows, during the snow-accumulation process, according to the laws presented in the distribution characteristics of each scattering component in the snow-free period and the snow-accumulation period, in both periods, the dihedral-angle scattering was the smallest, with a large difference from surface scattering and volume scattering; compared with the snow-free period, during the dry snow period, the difference between volume scattering and surface scattering increased and was parallel to the cross-section line trend; when there was snow cover, volume scattering was less than surface scattering, and when there was no snow cover, volume scattering was greater than surface scattering; compared with the snow-free period, the dihedral-angle scattering fluctuated less and had a smoother trend.

[0090] Meanwhile, using the optical remote sensing of the same period as a reference, in the image of Freeman polarization decomposition, the snow-covered area and non-snow-covered area during the dry snow period were selected, and the probability distribution density of each scattering characteristic was statistically analyzed (shown in Figure 7 and Figure 8 ).

[0091] As Figure 7 shows, in the snow-covered area, the distributions of the three scattering components, from small to large, were dihedral-angle scattering, volume scattering, and surface scattering; the dihedral-angle scattering was between -29 dB and -21 dB, mainly concentrated at -25 dB; the volume scattering was between -22 dB and -14 dB, mainly concentrated at -19 dB; the surface scattering was between -20 dB and -5 dB, with a plateau-shaped distribution characteristic and less concentrated values; it can be seen that the dihedral-angle scattering was more concentrated and had a high separability from both volume scattering and surface scattering; the volume scattering was generally less than the surface scattering, and due to the shallow snow cover, the volume scattering and surface scattering could not be directly separated, and there were many areas mixed together.

[0092] As Figure 8 shows, the distribution of the two scattering components, from small to large, was dihedral-angle scattering, surface scattering, and volume scattering; the dihedral-angle scattering was between -29 dB and -15 dB, mainly concentrated at -24 dB; the volume scattering was mainly concentrated at -8 dB and was generally larger than the surface scattering; the distribution of the surface scattering further expanded, with a distribution range between -20 dB and -1 dB.

[0093] Comparing Figure 7 and Figure 8, it is found that the volume scattering in the snow-covered area is less than that in the non-snow-covered area, and the surface scattering shows a plateau type in both the snow-covered area and the non-snow-covered area; from the analysis of the numerical range, there is a good separation between the dihedral angle scattering and the volume scattering and surface scattering, but both the dihedral angle scattering and the volume scattering are relatively concentrated.

[0094] Therefore, based on the laws of dihedral angle scattering and volume scattering, a difference index (i.e., snow cover index) is constructed from the dihedral angle scattering and the volume scattering:

[0095]

[0096] In the formula, FSCI _Vol_Dbl represents the snow cover degree result calculated from the volume scattering Freeman_Vol and the dihedral angle scattering Freeman_Dbl.

[0097] FSCI _Vol_Dbl Through the characteristics of volume scattering and surface scattering in the snow-covered area and the non-snow-covered area, the calculated range is between -1.0 and 0. Generally, the closer to 0, the higher the brightness value and the higher the snow cover degree.

[0098] S4. Statistically analyze the snow cover sample histogram based on the snow cover index, and obtain the recognition threshold according to the snow cover sample histogram.

[0099] On the basis of the above embodiments, in an optional embodiment of the present invention, step S4 includes steps S41 to S43.

[0100] S41. Obtain the snow cover index corresponding to a part of the snow-covered area as a sample, and statistically analyze the histogram of the sample according to the sample.

[0101] S42. Calculate the standard deviation and mean value of the sample according to the histogram.

[0102] S43. Obtain the recognition threshold according to the standard deviation and mean value.

[0103] In this embodiment, since the dihedral angle scattering, volume scattering, and surface scattering after Freeman polarization decomposition are related to the underlying surface, terrain, etc. in addition to being affected by snow cover, the threshold is not fixed. To determine the threshold, a part of the snow-covered area needs to be selected as a sample, the histogram is statistically analyzed, the degree of data dispersion is determined according to the standard deviation of the statistical histogram, and the threshold is determined near the mean value.

[0104] Specifically, when determining the threshold between snow and non-snow, by comparing the numerical values, it is possible to judge snow and non-snow and quickly extract the spatial distribution information of snow cover.

[0105] Since the dihedral angle scattering, volume scattering, and surface scattering after Freeman polarization decomposition are related to factors such as the underlying surface and terrain in addition to being affected by the presence or absence of snow cover, the threshold has a certain degree of uncertainty.

[0106] To set the threshold more accurately, in the embodiments of the present invention, a part of the snow-covered area in FSCI _Vol_Dbl is selected as a sample, the histogram is statistically analyzed, the degree of data dispersion is determined according to the standard deviation of the statistical histogram, and the threshold is determined near the mean value. According to the histogram, the data is normally distributed, the pixel mean is -0.161483, and the standard deviation is 0.044991. Since there are a small number of non-snow pixels in the snow pixels, in order to reduce the error, some smaller values are ignored. Through multiple experimental comparisons, the threshold -0.23 is selected, that is, when the FSCI _Vol_Dbl value is greater than -0.23, the pixel is snow-covered. In an optional embodiment, the mean plus the standard deviation is approximately equal to the threshold.

[0107] S5. According to the recognition threshold, judge the types of each pixel in the snow cover index to obtain the snow cover spatial distribution map of the target area. Among them, the types include snow and non-snow.

[0108] In some embodiments, step S5 includes steps S51 to S52.

[0109] S51. According to the threshold, judge whether the value of each pixel in the snow cover index is greater than the threshold.

[0110] S52. Assign the pixels with values larger than the threshold as snow pixels, and assign the pixels with values smaller than the threshold as non-snow pixels to obtain the snow cover spatial distribution map of the target area.

[0111] In this embodiment, by assigning the pixels larger than the threshold as snow pixels and the pixels smaller than the threshold as non-snow pixels, the snow cover spatial distribution map of the target area is obtained.

[0112] Specifically, the embodiments of the present invention reasonably and effectively utilize the features after Freeman full polarization decomposition, and select the polarization features that can better highlight the snow difference to construct an index, and quickly identify the snow through the threshold. The snow recognition method according to the embodiments of the present invention can quickly and accurately identify the snow-covered area according to the full polarization synthetic aperture radar image.

[0113] Under the premise of not requiring a large number of training samples, the embodiments of the present invention can quickly divide the snow / non-snow range by using a small amount of snow-covered areas, and the accuracy is equivalent to the result of supervised classification, which greatly improves the efficiency of mountain snow recognition and has certain reference significance for rapid snow recognition in large-scale mountains.

[0114] Embodiment 2

[0115] As Figure 9 shown, the embodiments of the present invention provide a snow recognition device based on Freeman decomposition, which includes:

[0116] An initial data acquisition module 1 is configured to acquire Radarsat-2 full-polarization image data of a target area.

[0117] A decomposition module 2 is configured to perform Freeman polarization decomposition on the Radarsat-2 full-polarization image data to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data. Among them, the polarization characteristics include the dihedral angle scattering characteristic Freeman_Dbl and the volume scattering characteristic Freeman_Vol.

[0118] An index construction module 3 is configured to construct a snow cover index FSCI of the Radarsat-2 full-polarization image data according to the dihedral angle scattering characteristic and the volume scattering characteristic _Vol_Dbl . Among them,

[0119] A threshold acquisition module 4 is configured to statistically analyze a snow cover sample histogram based on the snow cover index and obtain an identification threshold according to the snow cover sample histogram.

[0120] A snow cover identification module 5 is configured to determine the type of each pixel in the snow cover index according to the identification threshold to obtain a snow cover spatial distribution map of the target area. Among them, the types include snow cover and non-snow cover.

[0121] Based on the above embodiments, in an optional embodiment of the present invention, the decomposition module 2 includes:

[0122] A preprocessing unit is configured to preprocess the Radarsat-2 full-polarization image data. Among them, the preprocessing includes multi-look processing, polarization filtering, and geocoding through DEM data.

[0123] A matrix processing unit is configured to perform polarization matrix processing on the preprocessed Radarsat-2 full-polarization image data to obtain a polarization scattering matrix S and convert the polarization scattering matrix S into a polarization covariance matrix C.

[0124] A decomposition unit is configured to perform Freeman polarization decomposition according to the polarization covariance matrix C to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data.

[0125] Based on the above embodiments, in an optional embodiment of the present invention, the threshold acquisition module 4 includes:

[0126] A histogram acquisition unit is configured to obtain the snow cover index corresponding to a partial snow cover area as a sample and statistically analyze the histogram of the sample according to the sample.

[0127] A statistic calculation unit is configured to calculate the standard deviation and mean of the sample according to the histogram.

[0128] A threshold acquisition unit for acquiring an identification threshold according to the standard deviation and the mean value.

[0129] Based on the above embodiments, in an optional embodiment of the present invention, the initial data acquisition module 1 is specifically configured to: acquire Radarsat-2 full-polarization image data of a target area. Among them, the Radarsat-2 full-polarization image data includes four polarization modes: HH, HV, VV, and VH.

[0130] Based on the above embodiments, in an optional embodiment of the present invention, the snow cover identification module 5 includes:

[0131] A judgment unit for judging whether the value of each pixel in the snow cover index is greater than the threshold according to the threshold.

[0132] An identification unit for assigning pixels with values greater than the threshold as snow cover pixels and assigning pixels with values less than the threshold as non-snow cover pixels, so as to obtain a snow cover spatial distribution map of the target area.

[0133] Embodiment III

[0134] The embodiment of the present invention provides a snow cover identification device based on Freeman decomposition, which includes a processor, a memory, and a computer program stored in the memory. The computer program can be executed by the processor to implement the snow cover identification method based on Freeman decomposition described in any paragraph of Embodiment I.

[0135] Embodiment IV

[0136] The embodiment of the present invention provides a computer-readable storage medium, and the computer-readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the snow cover identification method based on Freeman decomposition described in any paragraph of Embodiment I.

[0137] In several embodiments provided by the embodiments of the present invention, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device and method embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0138] In addition, in each embodiment of the present invention, the various functional modules can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

[0139] If the described functions are implemented in the form of software function modules and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, an electronic device, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs, etc., which can store program codes. It should be noted that in this article, the term "including", "comprising", or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent to this process, method, article, or device. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the process, method, article, or device including the said element.

[0140] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. Snow recognition method based on Freeman decomposition, characterized in that, it includes: Obtain the Radarsat-2 full-polarization image data of the target area; Perform Freeman polarization decomposition on the Radarsat-2 full-polarization image data to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data; wherein, the polarization characteristics include dihedral scattering characteristics and volume scattering characteristics ; Construct a snow cover index for Radarsat-2 full-polarization image data based on the dihedral angle scattering characteristics and the volume scattering characteristics ; wherein, ; Statistical snow sample histogram according to the snow cover index, and obtain the recognition threshold according to the snow sample histogram; According to the recognition threshold, judge the types of each pixel in the snow cover index to obtain the snow cover spatial distribution map of the target area; wherein, the types include snow cover and non-snow cover; According to the Radarsat-2 full-polarization image data, perform Freeman polarization decomposition to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data, specifically including: Preprocess the Radarsat-2 full-polarization image data; wherein, the preprocessing includes multi-look processing, polarization filtering, and geocoding through DEM data; Perform polarization matrix processing on the preprocessed Radarsat-2 full-polarization image data to obtain the polarization scattering matrix S, and convert the polarization scattering matrix S into the polarization covariance matrix C; According to the polarization covariance matrix C, perform Freeman polarization decomposition to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data.

2. The snow recognition method based on Freeman decomposition according to claim 1, characterized in that, The polarization filtering is Refined Lee filtering; the DEM data is STRM4 DEM data; Obtain the Radarsat-2 full-polarization image data of the target area, specifically including: Obtain the Radarsat-2 full-polarization image data of the target area; wherein, the Radarsat-2 full-polarization image data includes four polarization modes: HH, HV, VV, and VH.

3. The snow recognition method based on Freeman decomposition according to claim 1, characterized in that, Statistical snow sample histogram according to the snow cover index, and obtain the recognition threshold according to the snow sample histogram, specifically including: Obtain the snow cover index corresponding to a part of the snow cover area as a sample, and statistically calculate the histogram of the sample according to the sample; According to the histogram, calculate the standard deviation and mean of the sample; According to the standard deviation and the mean, obtain the recognition threshold.

4. The snow recognition method based on Freeman decomposition according to claim 1, characterized in that, According to the recognition threshold, judge the types of each pixel in the snow cover index to obtain the snow cover spatial distribution map of the target area, specifically including: According to the threshold, judge whether the value of each pixel in the snow cover index is greater than the threshold; Assign the pixels with values larger than the threshold as snow cover pixels, and assign the pixels with values smaller than the threshold as non-snow cover pixels to obtain the snow cover spatial distribution map of the target area.

5. A snow recognition device based on Freeman decomposition, characterized in that, it includes: Initial data acquisition module, used to obtain the Radarsat-2 full-polarization image data of the target area; A decomposition module, configured to perform Freeman polarization decomposition on the Radarsat-2 full-polarization image data to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data; wherein the polarization characteristics include dihedral scattering characteristics and volume scattering characteristics ; An exponential component module for constructing a snow cover index of Radarsat-2 full polarization image data according to the dihedral angle scattering characteristics and the volume scattering characteristics wherein, ; A threshold acquisition module, configured to statistically analyze a snow cover sample histogram according to the snow cover index, and acquire an identification threshold according to the snow cover sample histogram; A snow cover identification module, configured to determine the types of each pixel in the snow cover index according to the identification threshold, so as to obtain a snow cover spatial distribution map of a target area; wherein, the types include snow cover and non-snow cover; The decomposition module includes: A preprocessing unit, configured to preprocess the Radarsat-2 full-polarization image data; wherein, the preprocessing includes multi-look processing, polarization filtering, and geocoding through DEM data; A matrix processing unit, configured to perform polarization matrix processing on the preprocessed Radarsat-2 full-polarization image data, acquire a polarization scattering matrix S, and convert the polarization scattering matrix S into a polarization covariance matrix C; A decomposition unit, configured to perform Freeman polarization decomposition according to the polarization covariance matrix C, so as to obtain the polarization characteristics of each pixel in the Radarsat-2 full-polarization image data.

6. The snow cover identification device based on Freeman decomposition according to claim 5, wherein, the threshold acquisition module includes: A histogram acquisition unit, configured to acquire the snow cover index corresponding to a partial snow cover area as a sample, and statistically analyze a histogram of the sample according to the sample; A statistic calculation unit, configured to calculate the standard deviation and mean of the sample according to the histogram; A threshold acquisition unit, configured to acquire the identification threshold according to the standard deviation and the mean.

7. A snow cover identification device based on Freeman decomposition, wherein, it includes a processor, a memory, and a computer program stored in the memory; the computer program can be executed by the processor to implement the snow cover identification method based on Freeman decomposition according to any one of claims 1 to 4.

8. A computer-readable storage medium, wherein, the computer-readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the snow cover identification method based on Freeman decomposition according to any one of claims 1 to 4.

Images