A flood inundation area full life cycle extraction method for a river basin scale

Abstract

The application discloses a flood inundation area full life cycle extraction method for a river basin scale, and relates to the field of remote sensing data processing. The method comprises the following steps: generating a first pseudo-color image according to first remote sensing image data of a target region, and generating a second pseudo-color image according to second remote sensing image data of the target region; adjusting the gray value of the second pseudo-color image on a target channel according to the gray value distribution information of the first pseudo-color image on the target channel, to obtain a third pseudo-color image; and analyzing the first pseudo-color image and the third pseudo-color image by using a semantic segmentation model, to obtain an inundation area in the target region. The application solves the technical problem that the inundation area cannot be accurately determined due to the fact that the remote sensing image data collected by different devices cannot be effectively utilized in the related art.

Description

Technical Field

[0001] This application relates to the field of remote sensing data processing, and more specifically, to a method for extracting the entire life cycle of flood inundation range at the watershed scale. Background Technology

[0002] In related technologies, the inability to effectively utilize remote sensing image data collected by different devices when determining the flooding range makes it impossible to accurately determine the flooding range of the target area.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This application provides a method for extracting the full life cycle of flood inundation range at the watershed scale, in order to at least solve the technical problem that the inundation range cannot be accurately determined due to the inability to effectively utilize remote sensing image data collected by different devices in related technologies.

[0005] According to one aspect of the embodiments of this application, a method for extracting the full life cycle of flood inundation range at the watershed scale is provided, comprising: generating a first pseudo-color image based on first remote sensing image data of a target area, and generating a second pseudo-color image based on second remote sensing image data of the target area, wherein the grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data, and the acquisition devices of the first and second remote sensing image data are different; adjusting the grayscale values ​​of the second pseudo-color image in the target channel based on the grayscale value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image, wherein the target channel is any one of the first and second pseudo-color images; and analyzing the first and third pseudo-color images using a semantic segmentation model to obtain the inundation range in the target area.

[0006] Optionally, the grayscale value distribution information includes a first cumulative distribution function of the grayscale values ​​of the first pseudo-color image in the target channel; adjusting the grayscale values ​​of points in the second pseudo-color image in the target channel according to the grayscale value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image includes: determining a second cumulative distribution function of the second pseudo-color image in the target channel; determining a target grayscale value corresponding to the initial grayscale value according to the first and second cumulative distribution functions, wherein the initial grayscale value is any pixel grayscale value of the second pseudo-color image in the target channel; and replacing the initial grayscale value with the target grayscale value to obtain the third pseudo-color image.

[0007] Optionally, determining the target gray value corresponding to the initial gray value based on the first cumulative distribution function and the second cumulative distribution function includes: determining the target cumulative probability corresponding to the initial gray value in the second cumulative distribution function based on the second cumulative distribution function; and determining the target gray value based on the target cumulative probability and the first cumulative distribution function, wherein the cumulative probability corresponding to the target gray value in the first cumulative distribution function is the target cumulative probability.

[0008] Optionally, the first remote sensing image data includes VV / VH dual-polarization backscattering remote sensing data, and the second remote sensing image data includes HH / HV dual-polarization backscattering remote sensing data.

[0009] Optionally, generating a first pseudo-color image based on first remote sensing image data of the target area includes: determining the grayscale value of the first pseudo-color image in a first channel based on VV single-polarization data in the first remote sensing image data; determining the grayscale value of the first pseudo-color image in a second channel based on VH single-polarization data in the first remote sensing image data; determining the grayscale value of the first pseudo-color image in a third channel based on the ratio of VV single-polarization data to VH single-polarization data; and generating the first pseudo-color image based on the grayscale values ​​of the first pseudo-color image in the first channel, the second channel, and the third channel.

[0010] Optionally, generating a second pseudo-color image based on second remote sensing image data of the target area includes: determining the grayscale value of the second pseudo-color image in the first channel based on HH single-polarization data in the second remote sensing image data; determining the grayscale value of the second pseudo-color image in the second channel based on HV single-polarization data in the second remote sensing image data; determining the grayscale value of the second pseudo-color image in the third channel based on the ratio of HH single-polarization data to HV single-polarization data; and generating the second pseudo-color image based on the grayscale values ​​of the second pseudo-color image in the first channel, the second channel, and the third channel.

[0011] Optionally, the method further includes: determining multiple detection times, and first and second remote sensing image data corresponding to the detection times; determining the inundation range corresponding to the detection times based on the first and second remote sensing image data corresponding to the detection times; and determining the inundation process of the target area based on the inundation range corresponding to each of the multiple detection times.

[0012] According to another aspect of the embodiments of this application, a device for extracting the full life cycle of flood inundation range at the watershed scale is also provided, comprising: a first processing module, used to generate a first pseudo-color image based on first remote sensing image data of a target area, and to generate a second pseudo-color image based on second remote sensing image data of the target area, wherein the grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data, and the polarization combination of the microwave remote sensing satellites corresponding to the first and second remote sensing image data are different; a second processing module, used to adjust the grayscale values ​​of the second pseudo-color image in the target channel based on the grayscale value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image, wherein the target channel is any one of the first and second pseudo-color images; and a third processing module, used to analyze the first pseudo-color image and the third pseudo-color image using a semantic segmentation model to obtain the inundation range in the target area.

[0013] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, wherein a program is stored in the non-volatile storage medium, wherein, when the program is running, it controls the device where the non-volatile storage medium is located to execute a method for extracting the full life cycle of flood inundation range at the watershed scale.

[0014] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, executes a method for extracting the full life cycle of flood inundation range at the watershed scale.

[0015] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of a method for extracting the full life cycle of flood inundation range at a watershed scale.

[0016] In this embodiment, a first pseudo-color image is generated based on first remote sensing image data of the target area, and a second pseudo-color image is generated based on second remote sensing image data of the target area. The grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, and the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data. The polarization combinations of the microwave remote sensing satellites corresponding to the first and second remote sensing image data are different. Based on the grayscale value distribution information of the first pseudo-color image in the target channel, the grayscale values ​​of the second pseudo-color image in the target channel are adjusted to obtain a third pseudo-color image. The target channel is the first... The method involves analyzing the first and third pseudo-color images using a semantic segmentation model to determine the flooding range in the target area. By identifying the pseudo-color images corresponding to different remote sensing image data and adjusting the grayscale values ​​of the second pseudo-color image in each channel based on the first pseudo-color image, the tones of the different pseudo-color images are aligned. This achieves the technical effect of comprehensively utilizing remote sensing image data collected by different devices to determine the flooding range, thereby solving the technical problem that the flooding range cannot be accurately determined due to the inability to effectively utilize remote sensing image data collected by different devices in related technologies. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 This is a schematic diagram of the structure of a computer terminal (or mobile device) according to an embodiment of this application;

[0019] Figure 2 This is a flowchart illustrating a method for extracting the entire lifecycle of flood inundation range at the watershed scale, according to an embodiment of this application.

[0020] Figure 3 This is a schematic diagram of the structure of a watershed-scale flood inundation range full life cycle extraction device provided in the embodiments of this application. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] To better understand the embodiments of this application, the technical terms involved in the embodiments of this application are explained below:

[0024] VV / VH dual-polarization backscattering remote sensing data: VV / VH dual-polarization backscattering remote sensing data is a type of dual-polarization remote sensing data acquired by synthetic aperture radar satellites. VV polarization represents radar signals transmitted and received vertically, while VH polarization represents radar signals transmitted and received horizontally. This type of data can be provided by mainstream remote sensing satellites such as Sentinel-1. It can stably acquire the backscattering characteristics of surface targets and clearly distinguish different land cover types such as water bodies, vegetation, and built-up areas. It is widely used in scenarios such as monitoring flood inundation areas and classifying land cover.

[0025] HH / HV Dual-Polarized Backscattering Remote Sensing Data: HH / HV dual-polarized backscattering remote sensing data is another type of dual-polarized remote sensing data acquired by synthetic aperture radar satellites. HH polarization represents radar signals transmitted and received horizontally, while HV polarization represents radar signals transmitted and received vertically. This type of data can be provided by the Gaofen-3 satellite, which has all-weather and all-time imaging capabilities. The combination of dual polarizations can enrich the dimensions of surface scattering information, effectively improve the accuracy of ground feature identification under complex surface backgrounds, and adapt to application needs such as watershed-scale flood dynamic monitoring.

[0026] Pseudo-color images: Pseudo-color images are visual images converted from single-band or multi-band remote sensing raw data into a three-channel format through numerical mapping, feature operations, etc. In this application, this type of image is generated by assigning grayscale values ​​to the corresponding channels based on the single-polarization components and component ratios of dual-polarization SAR remote sensing data. This transforms the ground cover scattering characteristics of remote sensing data into intuitive color distribution characteristics, enriches the feature expression dimensions of the data, adapts to the multi-channel input format of deep learning semantic segmentation models, and enhances the feature differentiation effect of ground cover types such as water bodies, vegetation, and built-up areas, providing a standardized image input foundation for accurate identification of inundation areas.

[0027] The Cumulative Distribution Function (CDF) is a core function used to quantitatively describe the statistical distribution of gray values ​​in a single channel of an image. Based on the statistical calculation of the gray values ​​of all pixels in the target channel of a pseudo-color image, this function can fully reflect the probability of gray values ​​accumulating from low to high. It can accurately characterize the overall distribution characteristics and statistical ranking relationship of all gray values ​​in the image channel. In this application, it is used to construct the gray value mapping relationship between different pseudo-color image channels, providing a standardized statistical basis for histogram matching and radiometric normalization processing of cross-sensor remote sensing images.

[0028] In recent years, remote sensing, with its advantages of wide-area coverage and periodic revisit, has become an important technical means for identifying the extent of flooding. Optical imagery and synthetic aperture radar (SAR) imagery are the most commonly used data sources for flood mapping. Optical imagery can provide long-term continuous observations and is suitable for constructing flood extent products and catalogs on long-term scales. SAR imagery has all-weather, day-night imaging capabilities and has been widely used in flood identification in recent years, gradually becoming the mainstream data source. Based on the above data, researchers have proposed a variety of flood identification methods, mainly including thresholding methods based on spectral indices or backscattering features, and deep learning models. Thresholding methods are widely used for emergency response and large-scale rapid mapping due to their high computational efficiency and ease of implementation; deep learning methods, on the other hand, have received increasing attention due to their superior identification performance compared to traditional methods.

[0029] Currently, the combination of deep learning models and SAR data has become the mainstream method for identifying flood inundation extent. Copernicus Emergency Management Services (CEMS) launched its Global Flood Monitoring (GFM) service in 2021, which is one of the most representative operational practices for identifying flood inundation extent in recent years: the system is global in scale and combines Sentinel-1 VV single-polarization SAR imagery with deep learning algorithms to achieve a fast link from image acquisition to flood product release.

[0030] However, existing methods still face several key limitations in watershed-scale applications, leading to a decline in model performance, mainly in the following two aspects:

[0031] Single-polarization SAR images have limited information dimensions. In complex terrain contexts such as urban areas and vegetated areas, flood signals in single-polarization SAR images are easily confused with mountain shadow effects and areas with different objects of the same spectrum, such as built-up areas, leading to false positives and false negatives.

[0032] The generalization ability across time phases, sensors, and large-scale applications is insufficient. Due to the limitation of SAR revisit period, capturing flood dynamics over a wide geographical area (such as a watershed scale) usually requires fusing SAR images from different sensors, regions, and time periods. Because these images differ in polarization, incident angle, and radiometric response, and manifest as tonal differences such as brightness / contrast at the image level, the performance of the trained model often degrades significantly when directly transferred to multi-temporal, multi-sensor SAR data.

[0033] To address the aforementioned issues, this application provides relevant solutions, which are detailed below.

[0034] According to an embodiment of this application, a method embodiment for extracting the full life cycle of flood inundation range at the watershed scale is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0035] The methods and embodiments provided in this application can be executed on mobile terminals, computer terminals, or similar computing devices. Figure 1 A hardware block diagram of a computer terminal (or mobile device) for implementing a method for extracting the full lifecycle of flood inundation extent at the watershed scale is shown. Figure 1 As shown, the computer terminal 10 (or mobile device 10) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0036] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be wholly or partially integrated into any other element within the computer terminal 10 (or mobile device 10). As involved in the embodiments of this application, the data processing circuit serves as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0037] The memory 104 can be used to store software programs and modules for application software, such as the program instructions / data storage device corresponding to the watershed-scale flood inundation range full life cycle extraction method in this embodiment of the application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the above-mentioned watershed-scale flood inundation range full life cycle extraction method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 (or mobile device 10) via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0038] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10 (or mobile device 10). In one example, the transmission device 106 includes a network interface controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a radio frequency (RF) module used for wireless communication with the Internet.

[0039] The display may be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10 (or mobile device 10).

[0040] Under the aforementioned operating environment, this application provides a method for extracting the entire lifecycle of flood inundation range at the watershed scale, such as... Figure 2 As shown, the method includes the following steps:

[0041] Step S202: Generate a first pseudo-color image based on the first remote sensing image data of the target area, and generate a second pseudo-color image based on the second remote sensing image data of the target area. The grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, and the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data. The polarization combination of the microwave remote sensing satellites corresponding to the first and second remote sensing image data is different.

[0042] In some embodiments of this application, the first remote sensing image data includes VV / VH dual-polarization backscattering remote sensing data, and the second remote sensing image data includes HH / HV dual-polarization backscattering remote sensing data. The first and second remote sensing image data can be remote sensing image data acquired by different satellites. For example, the first remote sensing image data can be Sentinel-1 imagery, and the second remote sensing image data can be Gaofen-3 SAR imagery. The target area includes the target watershed.

[0043] As an optional implementation, generating a first pseudo-color image based on first remote sensing image data of the target area includes: determining the grayscale value of the first pseudo-color image in a first channel based on VV single-polarization data in the first remote sensing image data; determining the grayscale value of the first pseudo-color image in a second channel based on VH single-polarization data in the first remote sensing image data; determining the grayscale value of the first pseudo-color image in a third channel based on the ratio of VV single-polarization data to VH single-polarization data; and generating the first pseudo-color image based on the grayscale values ​​of the first pseudo-color image in the first channel, the second channel, and the third channel.

[0044] Optionally, the first channel can be an R channel, the second channel can be a G channel, and the third channel can be a B channel.

[0045] In some embodiments of this application, the first remote sensing image data includes VV / VH dual-polarization backscattering remote sensing data. For each pixel in the remote sensing image within the target area, the grayscale value of that pixel in different channels of the first pseudo-color image can be generated based on the first remote sensing image data. For example, the VV single-polarization data of that pixel in the first remote sensing image data can be mapped to the grayscale value of the first channel of the first pseudo-color image, and the VH single-polarization data of that pixel in the first remote sensing image data can be mapped to the grayscale value of the second channel of the first pseudo-color image. The ratio of the VV single-polarization data to the VH single-polarization data of that pixel in the first remote sensing image data is calculated and mapped to the grayscale value of the third channel of the first pseudo-color image. Then, the first pseudo-color image is generated based on the grayscale values ​​of all pixels in the target area in the three channels.

[0046] In some embodiments of this application, generating a second pseudo-color image based on second remote sensing image data of a target area includes: determining the grayscale value of the second pseudo-color image in a first channel based on HH single-polarization data in the second remote sensing image data; determining the grayscale value of the second pseudo-color image in a second channel based on HV single-polarization data in the second remote sensing image data; determining the grayscale value of the second pseudo-color image in a third channel based on the ratio of HH single-polarization data to HV single-polarization data; and generating the second pseudo-color image based on the grayscale values ​​of the second pseudo-color image in the first channel, the second channel, and the third channel.

[0047] In some embodiments of this application, for each pixel in the remote sensing image within the target area, the grayscale value of that pixel in different channels of the second pseudo-color image can be generated based on the second remote sensing image data. For example, the HH single-polarization data of that pixel in the second remote sensing image data can be mapped to the grayscale value of the first channel of the second pseudo-color image, and the HV single-polarization data of that pixel in the second remote sensing image data can be mapped to the grayscale value of the second channel of the second pseudo-color image. Then, the ratio of the HH single-polarization data to the HV single-polarization data of that pixel in the second remote sensing image data is calculated, and this ratio is mapped to the grayscale value of the third channel of the second pseudo-color image. Then, the second pseudo-color image is generated based on the grayscale values ​​of all pixels in the target area in the three channels.

[0048] By converting remote sensing image data into pseudo-color images, dual-polarized single-channel scattering data can be expanded into three-channel feature data. This enriches the scattering feature representation of land features in the target area, enhances the feature differentiation of features such as water bodies and vegetation in built-up areas under mountain shadows, adapts to the multi-channel input format of semantic segmentation models, and improves the model's ability to extract features from inundated areas against complex surface backgrounds. Furthermore, it can fully exploit the feature information of dual-polarized scattering data, supplement the feature dimensions of the first pseudo-color image, weaken the differences in radiometric response between different sensor data, improve the model's generalization ability across sensor and temporal data processing, and reduce the false positive and false negative rates in inundation range identification.

[0049] Step S204: Based on the gray value distribution information of the first pseudo-color image in the target channel, adjust the gray value of the second pseudo-color image in the target channel to obtain the third pseudo-color image, wherein the target channel is any one of the first pseudo-color image and the second pseudo-color image;

[0050] Optionally, the grayscale distribution information of the second pseudo-color image in the target channel can also be used as a reference to adjust the grayscale values ​​of the first pseudo-color image in the target channel. The specific adjustment method is the same as when using the first pseudo-color image as a reference. The target channel can be the R channel, G channel, or B channel. During adjustment, only the grayscale values ​​of the second pseudo-color image in the R channel will be adjusted based on the grayscale distribution information of the first pseudo-color image in the R channel, the grayscale values ​​of the second pseudo-color image in the G channel will be adjusted based on the grayscale distribution information of the first pseudo-color image in the G channel, and the grayscale values ​​of the second pseudo-color image in the B channel will be adjusted based on the grayscale distribution information of the first pseudo-color image in the B channel.

[0051] In the technical solution provided in step S204, the grayscale value distribution information includes a first cumulative distribution function of the grayscale values ​​of the first pseudo-color image in the target channel; adjusting the grayscale values ​​of points in the second pseudo-color image in the target channel according to the grayscale value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image includes: determining a second cumulative distribution function of the second pseudo-color image in the target channel; determining a target grayscale value corresponding to the initial grayscale value according to the first cumulative distribution function and the second cumulative distribution function, wherein the initial grayscale value is any pixel grayscale value of the second pseudo-color image in the target channel; replacing the initial grayscale value with the target grayscale value to obtain the third pseudo-color image.

[0052] As an optional implementation, determining the target gray value corresponding to the initial gray value based on the first cumulative distribution function and the second cumulative distribution function includes: determining the target cumulative probability corresponding to the initial gray value in the second cumulative distribution function based on the second cumulative distribution function; and determining the target gray value based on the target cumulative probability and the first cumulative distribution function, wherein the cumulative probability corresponding to the target gray value in the first cumulative distribution function is the target cumulative probability.

[0053] In some embodiments of this application, the third pseudo-color image can be determined by histogram matching using the following process:

[0054] Step 1: Select the reference image and the image to be adjusted;

[0055] Optionally, a first pseudo-color image generated from the first remote sensing image data can be selected as a reference image for histogram matching. This reference image has standardized grayscale value statistical distribution characteristics. A second pseudo-color image generated from the second remote sensing image data can be selected as the image to be adjusted. This image to be adjusted has a tone and grayscale distribution that is inconsistent with the reference image. It is necessary to adjust the grayscale values ​​to align with the statistical distribution of the reference image.

[0056] Step 2: Calculate the cumulative distribution function of the target channel;

[0057] For any target channel corresponding to both the first and second pseudo-color images, the gray values ​​of all pixels in the target channel of the first pseudo-color image are completely traversed, and the occurrence frequency, cumulative frequency, and cumulative proportion of each gray value are statistically analyzed to generate a first cumulative distribution function that can completely represent the distribution pattern of gray values ​​in the target channel of the reference image. Simultaneously, the gray values ​​of all pixels in the target channel of the second pseudo-color image are completely traversed, and the occurrence frequency, cumulative frequency, and cumulative proportion of each gray value are calculated according to the same statistical rules as the first cumulative distribution function to generate a second cumulative distribution function that represents the original gray value distribution pattern of the target channel of the image to be adjusted.

[0058] Step 3: Determine the target cumulative probability of the pixel to be adjusted;

[0059] Randomly select the gray value of any pixel x from the target channel of the second pseudo-color image as the initial gray value. Substitute this initial gray value into the second cumulative distribution function obtained in step 2, and calculate the target cumulative probability corresponding to the initial gray value in the second cumulative distribution function through the function mapping relationship. p = F s ( x This cumulative probability can uniquely identify the statistical ranking and distribution characteristics of this gray value among all gray values ​​in the target channel of the image to be adjusted.

[0060] Step 4: Match the corresponding grayscale value of the reference image;

[0061] Substitute the target cumulative probability calculated in step 3 into the first cumulative distribution function obtained in step 2. In the mapping relationship between grayscale values ​​and cumulative probabilities in the first cumulative distribution function, find the grayscale value corresponding to the cumulative probability that is completely equal to the target cumulative probability value. That is, find the grayscale value y that satisfies Fr(y) = p in the first cumulative distribution function of the reference image, where Fr(y) represents the calculated cumulative probability of a point in the reference image in the first cumulative distribution function. Determine this grayscale value as the target grayscale value, ensuring that the cumulative probability corresponding to the target grayscale value in the first cumulative distribution function is consistent with the target cumulative probability corresponding to the initial grayscale value in the second cumulative distribution function. This establishes a stable monotonic mapping relationship between the image to be adjusted and the reference image.

[0062] Step 5: Complete the single-pixel single-channel grayscale value replacement;

[0063] The initial gray value selected in the target channel of the second pseudo-color image is replaced with the target gray value determined in step 4 to complete the gray value adjustment operation of the single pixel in the target channel, so that the adjusted gray value of the pixel conforms to the gray value statistical distribution characteristics of the target channel of the reference image.

[0064] Step 6: Complete the full-pixel adjustment of the target channel;

[0065] For all pixel grayscale values ​​in the target channel of the second pseudocolor image, repeat steps 3 to 5 in sequence to complete the matching calculation and replacement adjustment of each pixel grayscale value one by one until the grayscale values ​​of all pixels in the target channel of the second pseudocolor image are standardized and adjusted, so as to achieve the alignment of the overall grayscale value distribution of the target channel with the target channel of the reference image.

[0066] Step 7: Iterate through all channels to complete the grayscale value adjustment;

[0067] The remaining channels of the first pseudo-color image and the second pseudo-color image are used as new target channels in sequence. Steps 2 to 6 are repeated for each channel to complete the uniform adjustment of the gray values ​​of all channels of the second pseudo-color image. Finally, a third pseudo-color image with gray distribution and tonal features that perfectly match the reference image is obtained.

[0068] Through the above steps, accurate histogram matching and radiometric normalization processing of multi-source pseudo-color images based on the cumulative distribution function were achieved. By establishing a strict statistical mapping relationship of gray values ​​between the image to be adjusted and the reference image, the tonal, brightness, and contrast deviations caused by differences in incident angle radiometric response due to polarization methods in cross-sensor and cross-temporal remote sensing images are systematically eliminated. This ensures that pseudo-color images generated by different sensors have consistent gray-scale distribution characteristics at the statistical level, fundamentally solving the problem of insufficient model generalization ability when fusing multi-source remote sensing data. It significantly improves the stability and recognition accuracy of the semantic segmentation model in feature extraction in cross-sensor, cross-temporal, and large-scale watershed scenarios, effectively reducing the false detection and false negative rates of flood inundation range determination under complex surface backgrounds, and ensuring the reliability and consistency of the inundation range determination results.

[0069] Step S206: The semantic segmentation model is used to analyze the first pseudo-color image and the third pseudo-color image to obtain the flooding range in the target area.

[0070] In some embodiments of this application, the method further includes: determining multiple detection times and second remote sensing image data corresponding to the detection times; determining the inundation range corresponding to the detection times based on the first remote sensing image data and the second remote sensing image data corresponding to the detection times; and determining the inundation process of the target area based on the inundation range corresponding to each of the multiple detection times. This allows for the determination of the changes in the inundation range of the target area over multiple time periods. In other words, it determines the changes in the inundation range throughout the entire lifecycle of the inundation process.

[0071] Optionally, when determining the flooding range across multiple time periods, the first remote sensing image data can remain unchanged, and only the second remote sensing image data corresponding to different detection times can be determined. That is, the reference image data for the second remote sensing image data at each detection time is the first remote sensing image data.

[0072] In some embodiments of this application, to suppress interference from heterogeneous objects in complex surface backgrounds such as mountain shadows and built-up areas, representative negative samples covering typical scenes such as mountainous areas and built-up areas can be collected based on first remote sensing image data, and labeled ground truth data of typical flood-inundated areas within the target region can be compiled. Data augmentation processing is performed on the negative samples and labeled ground truth data to construct a training dataset with diverse sample types and a balance between positive and negative samples. This training dataset is used to train the DeepLabv3+ semantic segmentation model, which possesses multi-scale feature representation capabilities and robustness to complex backgrounds, improving the model's accuracy and anti-interference ability in identifying inundated areas under complex surface scenes. Multi-temporal first and second remote sensing image data covering the entire target watershed are input into the trained semantic segmentation model. The model inferences to obtain the flood inundation range corresponding to each detection time. Based on the inundation range at multiple detection times, the watershed-scale flood inundation process of the target region is determined, achieving a complete reconstruction of the flood inundation process.

[0073] To further verify the advantages of the method provided in the embodiments of this application over related technologies in determining the flood inundation range, some embodiments of this application select catastrophic flood events that have occurred in the target area as typical cases to verify the method provided in this application.

[0074] In some embodiments of this application, Sentinel-1 data of region A on the first date and Gaofen-3 data of region B on the second date are selected as validation areas. Ground truth data on the flood inundation extent within these areas is determined through manual interpretation. Based on the aforementioned ground truth data obtained through manual interpretation and the recognition results of the semantic segmentation model on the corresponding validation data, the Mean Intersection over Union (MIoU) is calculated to quantify the model's accuracy in identifying the flood inundation extent. The model trained using the method provided in the embodiments of this application maintains stable recognition performance on remote sensing image data acquired from different regions, different time phases, and different sensors, with its Mean Intersection over Union (MIoU) value consistently reaching approximately 90%.

[0075] The above verification results show that the semantic segmentation model trained by the method provided in the embodiments of this application has a stable ability to identify the flood inundation range on remote sensing image data acquired by different sensors in different regions and at different times. The average crossover ratio remains at a high level, which can effectively support the reconstruction needs of flood inundation processes at the watershed scale.

[0076] A first pseudo-color image is generated based on first remote sensing image data of the target area, and a second pseudo-color image is generated based on second remote sensing image data of the target area. The grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, and the grayscale values ​​of the second pseudo-color image in different channels are also determined based on the second remote sensing image data. The polarization combinations of the microwave remote sensing satellites corresponding to the first and second remote sensing image data are different. Based on the grayscale value distribution information of the first pseudo-color image in the target channel, the grayscale values ​​of the second pseudo-color image in the target channel are adjusted to obtain a third pseudo-color image, where the target channel is the first pseudo-color image. The method involves analyzing the first and third pseudo-color images using a semantic segmentation model to determine the flooding range in the target area. By identifying the pseudo-color images corresponding to different remote sensing image data and adjusting the grayscale values ​​of the second pseudo-color image in each channel based on the first pseudo-color image, the tones of the different pseudo-color images are aligned. This achieves the technical effect of comprehensively utilizing remote sensing image data collected by different devices to determine the flooding range, thereby solving the technical problem that the flooding range cannot be accurately determined due to the inability to effectively utilize remote sensing image data collected by different devices in related technologies.

[0077] This application provides a device for improving the entire life cycle of flood inundation range at the watershed scale. Figure 3 This is a schematic diagram of the device. From Figure 3 As can be seen from the diagram, the device includes: a first processing module 50, used to generate a first pseudo-color image based on first remote sensing image data of the target area, and a second pseudo-color image based on second remote sensing image data of the target area, wherein the grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data, and the polarization combination methods of the microwave remote sensing satellites corresponding to the first and second remote sensing image data are different; a second processing module 52, used to adjust the grayscale values ​​of the second pseudo-color image in the target channel based on the grayscale value distribution information of the first pseudo-color image in the target channel, to obtain a third pseudo-color image, wherein the target channel is any one of the first and second pseudo-color images; and a third processing module 54, used to analyze the first and third pseudo-color images using a semantic segmentation model to obtain the flooding range in the target area.

[0078] In some embodiments of this application, the first remote sensing image data includes VV / VH dual-polarization backscattering remote sensing data, and the second remote sensing image data includes HH / HV dual-polarization backscattering remote sensing data.

[0079] In some embodiments of this application, the step of the first processing module 50 generating a first pseudo-color image based on the first remote sensing image data of the target area includes: determining the grayscale value of the first pseudo-color image in the first channel based on the VV single-polarization data in the first remote sensing image data; determining the grayscale value of the first pseudo-color image in the second channel based on the VH single-polarization data in the first remote sensing image data; determining the grayscale value of the first pseudo-color image in the third channel based on the ratio of the VV single-polarization data to the VH single-polarization data; and generating the first pseudo-color image based on the grayscale values ​​of the first pseudo-color image in the first channel, the second channel, and the third channel.

[0080] In some embodiments of this application, the step of the first processing module 50 generating a second pseudo-color image based on the second remote sensing image data of the target area includes: determining the grayscale value of the second pseudo-color image in the first channel based on the HH single-polarization data in the second remote sensing image data; determining the grayscale value of the second pseudo-color image in the second channel based on the HV single-polarization data in the second remote sensing image data; determining the grayscale value of the second pseudo-color image in the third channel based on the ratio of the HH single-polarization data to the HV single-polarization data; and generating the second pseudo-color image based on the grayscale values ​​of the second pseudo-color image in the first channel, the second channel, and the third channel.

[0081] In some embodiments of this application, the grayscale value distribution information includes a first cumulative distribution function of the grayscale values ​​of the first pseudo-color image in the target channel; the step of the second processing module 52 adjusting the grayscale values ​​of points in the second pseudo-color image in the target channel according to the grayscale value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image includes: determining a second cumulative distribution function of the second pseudo-color image in the target channel; determining a target grayscale value corresponding to the initial grayscale value according to the first cumulative distribution function and the second cumulative distribution function, wherein the initial grayscale value is any pixel grayscale value of the second pseudo-color image in the target channel; replacing the initial grayscale value with the target grayscale value to obtain the third pseudo-color image.

[0082] In some embodiments of this application, the step of the second processing module 52 determining the target gray value corresponding to the initial gray value based on the first cumulative distribution function and the second cumulative distribution function includes: determining the target cumulative probability corresponding to the initial gray value in the second cumulative distribution function based on the second cumulative distribution function; determining the target gray value based on the target cumulative probability and the first cumulative distribution function, wherein the cumulative probability corresponding to the target gray value in the first cumulative distribution function is the target cumulative probability.

[0083] In some embodiments of this application, the device for determining the full life cycle of flood inundation range at the watershed scale is further configured to: determine multiple detection times and the second remote sensing image data corresponding to the detection times; determine the inundation range corresponding to the detection times based on the first remote sensing impact data and the second remote sensing image data corresponding to the detection times; and determine the inundation process of the target area based on the inundation range corresponding to each of the multiple detection times.

[0084] It should be noted that the modules in the above-mentioned watershed-scale flood inundation range full life cycle improvement device can be program modules (e.g., a set of program instructions to implement a certain function) or hardware modules. For the latter, it can be expressed in the following forms, but is not limited to them: each of the above modules is expressed as a processor, or the functions of each of the above modules are implemented by a processor.

[0085] According to an embodiment of this application, a non-volatile storage medium is also provided, which stores a program. During program execution, the device containing the non-volatile storage medium executes the following watershed-scale flood inundation range full lifecycle extraction method: A first pseudo-color image is generated based on first remote sensing image data of the target area, and a second pseudo-color image is generated based on second remote sensing image data of the target area. The grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, and the grayscale values ​​of the second pseudo-color image in different channels are also determined based on the second remote sensing image data. The polarization combination of the microwave remote sensing satellites corresponding to the first and second remote sensing image data is different. Based on the grayscale value distribution information of the first pseudo-color image in the target channel, the grayscale values ​​of the second pseudo-color image in the target channel are adjusted to obtain a third pseudo-color image. The target channel is any one of the first and second pseudo-color images. A semantic segmentation model is used to analyze the first and third pseudo-color images to obtain the inundation range in the target area.

[0086] According to an embodiment of this application, an electronic device is also provided, including: a memory and a processor. The processor is used to run a program stored in the memory, wherein the program executes the following method for extracting the full life cycle of flood inundation range at a watershed scale: generating a first pseudo-color image based on first remote sensing image data of the target area, and generating a second pseudo-color image based on second remote sensing image data of the target area, wherein the grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data, and the polarization combination of the microwave remote sensing satellites corresponding to the first and second remote sensing image data are different; adjusting the grayscale values ​​of the second pseudo-color image in the target channel based on the grayscale value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image, wherein the target channel is any one of the first and second pseudo-color images; and analyzing the first and third pseudo-color images using a semantic segmentation model to obtain the inundation range in the target area.

[0087] According to an embodiment of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the following steps of a method for extracting the entire lifecycle of flood inundation range at a watershed scale: generating a first pseudo-color image based on first remote sensing image data of the target area, and generating a second pseudo-color image based on second remote sensing image data of the target area, wherein the grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data, and the polarization combination of the microwave remote sensing satellites corresponding to the first and second remote sensing image data are different; adjusting the grayscale values ​​of the second pseudo-color image in the target channel based on the grayscale value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image, wherein the target channel is any one of the first and second pseudo-color images; and analyzing the first and third pseudo-color images using a semantic segmentation model to obtain the inundation range in the target area.

[0088] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0091] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0092] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to related technologies, or all or part of the 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 to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0093] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for extracting the entire life cycle of flood inundation range at the watershed scale, characterized in that, include: A first pseudo-color image is generated based on the first remote sensing image data of the target area, and a second pseudo-color image is generated based on the second remote sensing image data of the target area. The grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, and the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data. The polarization combination methods of the microwave remote sensing satellites corresponding to the first remote sensing image data and the second remote sensing image data are different. Based on the grayscale value distribution information of the first pseudocolor image in the target channel, the grayscale value of the second pseudocolor image in the target channel is adjusted to obtain a third pseudocolor image, wherein the target channel is any one of the first pseudocolor image and the second pseudocolor image; A semantic segmentation model is used to analyze the first pseudo-color image and the third pseudo-color image to obtain the flooding range in the target region.

2. The method for extracting the entire life cycle of flood inundation range at the watershed scale according to claim 1, characterized in that, The grayscale value distribution information includes a first cumulative distribution function of the grayscale values ​​of the first pseudo-color image in the target channel; based on the grayscale value distribution information of the first pseudo-color image in the target channel, adjusting the grayscale values ​​of points in the second pseudo-color image in the target channel to obtain a third pseudo-color image includes: Determine the second cumulative distribution function of the second pseudo-color image on the target channel; Based on the first cumulative distribution function and the second cumulative distribution function, the target gray value corresponding to the initial gray value is determined, wherein the initial gray value is any pixel gray value of the second pseudo-color image in the target channel; The initial grayscale value is replaced with the target grayscale value to obtain the third pseudo-color image.

3. The method for extracting the entire life cycle of flood inundation range at the watershed scale according to claim 2, characterized in that, Based on the first cumulative distribution function and the second cumulative distribution function, the target gray value corresponding to the initial gray value is determined as follows: Based on the second cumulative distribution function, determine the target cumulative probability corresponding to the initial gray value in the second cumulative distribution function; The target gray value is determined based on the target cumulative probability and the first cumulative distribution function, wherein the cumulative probability corresponding to the target gray value in the first cumulative distribution function is the target cumulative probability.

4. The method for extracting the entire life cycle of flood inundation range at the watershed scale according to claim 1, characterized in that, The first remote sensing image data includes VV / VH dual-polarization backscattering remote sensing data, and the second remote sensing image data includes HH / HV dual-polarization backscattering remote sensing data.

5. The method for extracting the entire life cycle of flood inundation range at the watershed scale according to claim 4, characterized in that, Generating the first pseudo-color image based on the first remote sensing image data of the target area includes: The grayscale value of the first pseudocolor image in the first channel is determined based on the VV single-polarization data in the first remote sensing image data; The grayscale value of the first pseudocolor image in the second channel is determined based on the VH single-polarization data in the first remote sensing image data; The grayscale value of the first pseudo-color image in the third channel is determined based on the ratio of the VV single-polarization data to the VH single-polarization data; The first pseudo-color image is generated based on the gray values ​​of the first pseudo-color image in the first channel, the second channel, and the third channel.

6. The method for extracting the entire life cycle of flood inundation range at the watershed scale according to claim 4, characterized in that, Generating a second pseudo-color image based on second remote sensing image data of the target area includes: The grayscale value of the second pseudocolor image in the first channel is determined based on the HH single-polarization data in the second remote sensing image data; The grayscale value of the second pseudocolor image in the second channel is determined based on the HV single-polarization data in the second remote sensing image data; The grayscale value of the second pseudocolor image in the third channel is determined based on the ratio of the HH single-polarization data to the HV single-polarization data; The second pseudo-color image is generated based on the gray values ​​of the second pseudo-color image in the first channel, the second channel, and the third channel.

7. The method for extracting the entire life cycle of flood inundation range at the watershed scale according to claim 1, characterized in that, The method further includes: Multiple detection times are determined, along with the second remote sensing image data corresponding to each detection time; Based on the first remote sensing image data and the second remote sensing image data corresponding to the detection time, the flooding range corresponding to the detection time is determined; The flooding process of the target area is determined based on the flooding range corresponding to each of the plurality of detection times.

8. A device for extracting the entire life cycle of flood inundation range at the watershed scale, characterized in that, include: The first processing module is used to generate a first pseudo-color image based on the first remote sensing image data of the target area, and to generate a second pseudo-color image based on the second remote sensing image data of the target area. The grayscale values ​​of the first pseudo-color image in different channels are determined based on the first remote sensing image data, and the grayscale values ​​of the second pseudo-color image in different channels are determined based on the second remote sensing image data. The polarization combination methods of the microwave remote sensing satellites corresponding to the first remote sensing image data and the second remote sensing image data are different. The second processing module is used to adjust the gray values ​​of the second pseudo-color image in the target channel according to the gray value distribution information of the first pseudo-color image in the target channel to obtain a third pseudo-color image, wherein the target channel is any one of the first pseudo-color image and the second pseudo-color image; The third processing module is used to analyze the first pseudo-color image and the third pseudo-color image using a semantic segmentation model to obtain the flooding range in the target region.

9. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores a program, wherein when the program is executed, it controls the device containing the non-volatile storage medium to execute the full life-cycle extraction method for flood inundation range at the watershed scale as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, executes the watershed-scale flood inundation range full lifecycle extraction method as described in any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for extracting the full life cycle of flood inundation range at the watershed scale as described in any one of claims 1 to 7.

Images