Large-area photovoltaic land classification method and system based on joint decision tree and deep learning

Abstract

The invention discloses a large-area photovoltaic land classification method combining a decision tree and deep learning, and belongs to the field of machine learning and remote sensing intelligent mapping. In order to solve the problem that single traditional machine learning and deep learning are difficult to consider the precision and speed of large-area remote sensing image classification, a global representative area sample set is made based on remote sensing images, a photovoltaic land use index is constructed, and a two-step classification method is adopted: the first step, a decision tree classifier is constructed based on the photovoltaic land use index and other characteristics, and a decision tree classifier is constructed based on the photovoltaic land use index and other characteristics; rapid classification of large-area photovoltaic land is realized on a GEE remote sensing cloud platform, and a rough classification area of the photovoltaic land is generated after a classification result is processed by post-processing, grid-vector conversion, a buffer area and the like; and a second step of adding an auxiliary branch and an index space attention module in the baseline ResUNet + + model, constructing a photovoltaic land classification model of multi-scale fusion photovoltaic land features, and accurately identifying the photovoltaic land from the rough classification area based on the model, thereby realizing rapid and accurate drawing of the large-area photovoltaic land.

Description

Technical Field

[0001] The present invention belongs to the field of machine learning and remote sensing intelligent mapping, and in particular relates to a large-area photovoltaic land classification method and system combining a decision tree and deep learning. Background Art

[0002] Photovoltaic land (PVL) refers to a new land use method that installs photovoltaic panels on underlying land use / cover backgrounds such as bare land (Gobi, bare land, saline-alkali land, etc.), vegetation (grassland, woodland, farmland, etc.), water surface and roof. Carrying out large-area photovoltaic land monitoring can provide data support for photovoltaic project site selection, photovoltaic power generation prediction, the impact of photovoltaic project construction on the ecological environment, and the formulation of green energy policies.

[0003] Existing PV data are mainly released by energy and related think tank organizations, with statistical units in megawatts (GW) and time scales in years. The advantage of this type of data is that it can reflect the PV installed capacity of a country or region and the location information of large PV facilities. The disadvantage is that it lacks information such as area and the location information of small PV facilities. Therefore, this type of data has limited reference value for dynamic monitoring of PV land use.

[0004] Satellite remote sensing is the most effective means to achieve dynamic monitoring of large-area photovoltaic land. Although there are many successful cases of using traditional machine learning or deep learning methods to identify large-area photovoltaic land from satellite or airborne remote sensing images, current research still lacks a global fine sample set that takes into account the underlying land use / cover background type, photovoltaic array installation inclination, geographical location, etc., and there is little intervention in domain knowledge, and it is difficult to take into account the speed and accuracy of large-area PVL monitoring. Therefore, on the basis of constructing a global representative regional sample set, it is urgent to develop a large-area photovoltaic land rapid and precise identification method with spatiotemporal migration capabilities based on open source satellite images. Summary of the invention

[0005] The purpose of the present invention is to solve the problem in the prior art that it is difficult to achieve both rapidity and accuracy in remote sensing classification of large-area photovoltaic land, and to provide a large-area photovoltaic land classification method and system combining decision trees and deep learning.

[0006] In order to achieve the above-mentioned invention object, the present invention specifically adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a large-area photovoltaic land classification method combining decision trees and deep learning, which comprises the following steps:

[0008] S1. Select data of representative photovoltaic land areas from the global photovoltaic land location dataset and classification dataset, and construct a photovoltaic land sample dataset containing remote sensing images, geographic coordinates, imaging time, underlying land cover type, and photovoltaic array arrangement mode;

[0009] S2, inputting the photovoltaic land sample data set into the constructed decision tree classifier for rough classification, and obtaining a rough segmentation result in the form of a binary raster map;

[0010] S3, post-processing the coarse segmentation result, setting a buffer zone of a preset size around the post-processed coarse segmentation result, framing an outer envelope rectangle outside the buffer zone, and forming a coarse classification area image of the photovoltaic land by the pixels inside the outer envelope rectangle in the remote sensing image;

[0011] S4. Generate a real label for each photovoltaic land rough classification area image through human-computer interaction, and slice the photovoltaic land rough classification area image and the real label of photovoltaic land, so that the photovoltaic land classification data set is formed by the sliced ​​photovoltaic land rough classification area image and the real label of photovoltaic land;

[0012] S5. Train a photovoltaic land classification model that integrates multi-scale photovoltaic land features on the photovoltaic land classification dataset;

[0013] S6. Input the image of the coarse classification area of ​​the photovoltaic land to be classified into the trained photovoltaic land classification model to obtain the segmentation result of the image of the coarse classification area of ​​the photovoltaic land, so as to realize the fine identification of the photovoltaic land.

[0014] Based on the above scheme, each step can be implemented in the following preferred specific manner.

[0015] As a preferred embodiment of the above-mentioned first aspect, in step S2, the classification limitation conditions of the decision tree classifier are divided into photovoltaic land index conditions and other conditions, the photovoltaic land index conditions are the first photovoltaic land index, the second photovoltaic land index and the blue-red band ratio, and the other conditions are slope, altitude, normalized building index and reflectivity of the first short-wave infrared band.

[0016] As a preferred embodiment of the first aspect, the specific calculation method of the first photovoltaic land index PVLI1, the second photovoltaic land index PVLI2, the blue-red band ratio BDRI, and the normalized building index NDBI is:

[0017]

[0018]

[0019] Among them, ρ Blue ,ρ Red ,ρ NIR ,ρSWIR1 ,ρ SWIR2 They represent the reflectivity of the blue band, red band, near-infrared band, first short-wave infrared band, and second short-wave infrared band respectively.

[0020] As a preferred embodiment of the first aspect, in step S2, the basis for coarse classification of the photovoltaic land sample data set is set as follows: simultaneously satisfying the first photovoltaic land index PVLI1∈[0.75,3.65], the second photovoltaic land index PVLI1∈[0.30,3.60], the blue-red band ratio BDRI∈[0.56,1.21], the slope SLOPE∈[0,20°], the altitude ALTITUDE∈[0,3500m], the normalized building index NDBI∈[-0.03,0.46] and the reflectivity ρ of the first short-wave infrared band SWIR1 ∈[0.07,0.58].

[0021] As a preferred embodiment of the above-mentioned first aspect, in step S3, the specific process of post-processing the coarse segmentation result is: first, the coarse segmentation result is subjected to spatial mean filtering, then the spatial mean filtering result is subjected to reclassification and denoising, and then the reclassification and denoising result in the form of a binary raster image is vectorized to obtain a vectorized binary image, and finally, all elements in the vectorized binary image whose area is greater than a preset area threshold are removed to obtain a post-processed coarse segmentation result.

[0022] As a preferred embodiment of the above-mentioned first aspect, the photovoltaic land classification model is a main-sub dual-branch structure, the main branch includes an input module, an encoder, a void space pyramid pooling module, a decoder and an output module, the sub-branch includes an input module, an encoder and five exponential space attention modules, wherein the encoder of the main branch, the encoder of the sub-branch and the decoder of the main branch each contain four residual blocks, the main branch is a U-shaped symmetrical structure, each residual block of the main branch decoder is jump-connected with each residual block at the same level of the main branch encoder, the input and output of each residual block of the sub-branch encoder are densely connected, the exponential feature space weight generated by each exponential space attention module in the sub-branch is densely connected with each residual block at the same level of the main branch encoder, the exponential feature space weight generated by each exponential space attention module in the sub-branch is densely connected with the exponential feature space weight generated by the next exponential space attention module, and then densely connected with each residual block at the same level of the main branch decoder.

[0023] As a preferred embodiment of the above-mentioned first aspect, in the photovoltaic land classification model of step S5, the specific processing flow is: based on the image of the coarse classification area of ​​the photovoltaic land after slicing, a multispectral image and an index image are generated, the multispectral image is input into the input module of the main branch to obtain a first spectral feature map, the index image is input into the input module of the sub-branch to obtain a first index feature map, the first index feature map is input into the first index space attention module of the sub-branch to obtain a first index feature space weight matrix, the first index feature space weight matrix and the first spectral feature map are multiplied to obtain a first fused feature map; the first fused feature map is input into the first residual block of the main branch encoder to obtain a second spectral feature map, and the first index is multiplied into the first residual block of the main branch encoder to obtain a second spectral feature map. The feature map is input into the first residual block of the sub-branch encoder to obtain a second exponential feature map, the first exponential feature map and the second exponential feature map are densely connected to obtain a first densely connected feature map, the first densely connected feature map is input into the second exponential space attention module of the sub-branch to obtain a second exponential feature space weight matrix, the second exponential feature space weight matrix and the second spectral feature map are multiplied to obtain a second fused feature map; the second fused feature map is input into the second residual block of the main branch encoder to obtain a third spectral feature map, the first densely connected feature map is input into the second residual block of the sub-branch encoder to obtain a third exponential feature map, the third exponential feature map and the first densely connected feature map are multiplied to obtain a third spectral feature map. The rows are densely connected to obtain a second densely connected feature map, the second densely connected feature map is input into the third exponential space attention module of the sub-branch to obtain a third exponential feature space weight matrix, the third exponential feature space weight matrix and the third spectral feature map are multiplied to obtain a third fused feature map; the third fused feature map is input into the third residual block of the main branch encoder to obtain a fourth spectral feature map, the second densely connected feature map is input into the third residual block of the sub-branch encoder to obtain a fourth exponential feature map, the fourth exponential feature map and the second densely connected feature map are densely connected to obtain a third densely connected feature map, and the third densely connected feature map is input into the fourth exponential space attention module of the sub-branch. Obtain a fourth exponential feature space weight matrix, multiply the fourth exponential feature space weight matrix and the fourth spectral feature map to obtain a fourth fused feature map; input the fourth fused feature map into the fourth residual block of the main branch encoder to obtain a fifth spectral feature map, input the third densely connected feature map into the fourth residual block of the sub-branch encoder to obtain a fifth exponential feature map, densely connect the fifth exponential feature map and the third densely connected feature map to obtain a fourth densely connected feature map, input the fourth densely connected feature map into the fifth exponential space attention module of the sub-branch to obtain a fifth exponential feature space weight matrix, and multiply the fifth exponential feature space weight matrix and the fifth spectral feature map to obtain a fifth fused feature map;The fifth fused feature map is input into the atrous space pyramid pooling module to obtain the atrous space pooling feature map, the atrous space pooling feature map is multiplied by the fifth exponential feature space weight matrix to obtain the sixth fused feature map, the sixth fused feature map is input into the fourth residual block of the main branch decoder to obtain the sixth spectral feature map, the fifth exponential feature space weight matrix and the fourth exponential feature space weight matrix are densely connected to obtain the first dense connection weight matrix, the sixth spectral feature map is multiplied by the first dense connection weight matrix to obtain the seventh fused feature map; the seventh fused feature map is input into the third residual block of the main branch decoder to obtain the seventh spectral feature map, the first dense connection weight matrix and the third exponential feature space weight matrix are densely connected to obtain the second dense connection weight matrix, the seventh spectral feature map and the second dense connection weight matrix are densely connected The eighth fusion feature map is obtained by multiplying the weight matrix; the eighth fusion feature map is input into the second residual block of the main branch decoder to obtain the eighth spectral feature map, the second dense connection weight matrix and the second exponential feature space weight matrix are densely connected to obtain the third dense connection weight matrix, and the eighth spectral feature map and the third dense connection weight matrix are multiplied to obtain the ninth fusion feature map; the ninth fusion feature map is input into the first residual block of the main branch decoder to obtain the ninth spectral feature map, the third dense connection weight matrix and the first exponential feature space weight matrix are densely connected to obtain the fourth dense connection weight matrix, and the ninth spectral feature map and the fourth dense connection weight matrix are multiplied to obtain the tenth fusion feature map; the tenth fusion feature map is input into the output module of the main branch to obtain the segmentation result output by the photovoltaic land classification model. ;

[0024] As a preferred embodiment of the first aspect, the multispectral image is obtained in the following manner:

[0025] In the image of the rough classified area of ​​photovoltaic land after slicing, the images of the blue band, green band, red band and the first short-wave infrared band are taken for band synthesis to form a multispectral image.

[0026] As a preferred embodiment of the first aspect, the index image is obtained in the following manner:

[0027] The images of the red band, near-infrared band, first short-wave infrared band and second short-wave infrared band are taken from the image of the coarse classification area of ​​photovoltaic land after slicing, and are calculated according to the calculation formula of the first photovoltaic land index to obtain the first photovoltaic land index image; the images of the blue band, red band, near-infrared band, first short-wave infrared band and second short-wave infrared band are taken from the image of the coarse classification area of ​​photovoltaic land after slicing, and are calculated according to the calculation formula of the second photovoltaic land index to obtain the second photovoltaic land index image; the first photovoltaic land index image and the second photovoltaic land index image are band synthesized to form an index image.

[0028] In a second aspect, the present invention provides a large-area photovoltaic land classification system combining decision trees and deep learning, comprising:

[0029] The data acquisition module is used to select data of representative photovoltaic land areas from the global photovoltaic land location data set and classification data set, and construct a photovoltaic land sample data set containing remote sensing images, geographic coordinates, imaging time, underlying land cover type, and photovoltaic array arrangement mode;

[0030] A coarse classification module is used to input the photovoltaic land sample data set into the constructed decision tree classifier for coarse classification, and obtain a coarse segmentation result in the form of a binary raster map;

[0031] A post-processing module is used to post-process the rough segmentation results, set a buffer zone of a preset size around the post-processed rough segmentation results, frame an outer envelope rectangle outside the buffer zone, and form a rough classification area image of photovoltaic land from the pixels inside the outer envelope rectangle in the remote sensing image;

[0032] The dataset construction module is used to generate a real photovoltaic land label for each photovoltaic land coarse classification area image through human-computer interaction, and to slice the photovoltaic land coarse classification area image and the photovoltaic land real label, so that the photovoltaic land classification dataset is formed by the sliced ​​photovoltaic land coarse classification area image and the photovoltaic land real label;

[0033] Model training module, used to train a photovoltaic land classification model that integrates photovoltaic land features at multiple scales on the photovoltaic land classification dataset;

[0034] The result acquisition module is used to input the coarse classification area image of the photovoltaic land to be classified into the trained photovoltaic land classification model, obtain the segmentation result of the coarse classification area image of the photovoltaic land, and realize the fine identification of the photovoltaic land.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] In view of the difficulty in balancing speed and accuracy in remote sensing classification of photovoltaic land in large areas, the present invention proposes a large-area photovoltaic land classification method based on the construction of a globally representative photovoltaic land sample data set. Based on the rough classification results of photovoltaic land obtained by using a decision tree classifier, the present invention combines four post-processing techniques, namely spatial mean filtering, reclassification denoising, area threshold denoising, and grid-vector conversion, to correct the rough classification results, and then creates a buffer zone and an outer envelope rectangle to delineate the PVL rough classification area, which greatly reduces the amount of data that needs to be processed by deep learning classification.

[0037] The present invention also constructs a photovoltaic land classification model PVLI-FSSM that integrates the photovoltaic land index. After training and hyperparameter optimization, the model achieves higher segmentation accuracy than the existing deep learning model. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Schematic diagram of the steps of the method of the present invention;

[0039] Figure 2 It is a schematic diagram of the framework of the method of the present invention;

[0040] Figure 3 It is a structural schematic diagram of the photovoltaic land classification model of the present invention;

[0041] Figure 4 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned purpose, features and advantages of the present invention more obvious and easy to understand, the specific implementation mode of the present invention is described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below. The technical features in each embodiment of the present invention can be combined accordingly without conflicting with each other.

[0043] In the description of the present invention, it should be understood that the terms "first" and "second" are only used for the purpose of distinguishing descriptions, and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features.

[0044] like Figure 1 As shown, in a preferred implementation of the present invention, the above-mentioned large-area photovoltaic land classification method combining decision tree and deep learning includes the following steps S1 to S6. The specific implementation process is described in detail below.

[0045] S1. Data of representative photovoltaic land (PVL) areas are selected from the global photovoltaic land location dataset and classification dataset to construct a photovoltaic land sample dataset containing remote sensing images, geographic coordinates, imaging time, underlying land cover type and photovoltaic array arrangement.

[0046] It should be noted that the location dataset of global photovoltaic land mentioned in the present invention refers to various points where photovoltaic land is measured. Such point information mainly comes from the installation statistics of the National Energy Administration of various countries and the measured data disclosed by various scientific research institutions. Each measured point will save four information of the main underlying land cover type, photovoltaic array arrangement mode and corresponding geographic coordinates and sampling time when recording. According to the geographic coordinates and sampling time information, the corresponding remote sensing image can be collected, and the above information can be attached to the remote sensing image to construct it. In addition, the above-mentioned classification dataset of global photovoltaic land can adopt the existing global photovoltaic land dataset, or it can be self-collected data. For example, the underlying land cover type, photovoltaic array arrangement mode and corresponding geographic coordinates and sampling time of photovoltaic land in different regions in the world can directly adopt the global photovoltaic land vector dataset (GPVI). GPVI is the core global photovoltaic land vector data product recognized by Nature journal for geographic information science, which contains the recognition results of global photovoltaic land from 2016 to 2019, and the positioning accuracy exceeds 85%. The underlying land cover type, photovoltaic array arrangement, and corresponding geographic coordinates and sampling time data of photovoltaic land in different regions around the world can be extracted from the GPVI data.

[0047] It should also be noted that since the method of the present invention requires a large amount of photovoltaic land data worldwide to maximize the versatility of the photovoltaic land classification model, and there is spatial heterogeneity in different regions around the world, the photovoltaic land sample data set should also cover the global scope as much as possible. In actual use, if only a large-scale photovoltaic land is to be quickly and accurately identified in a specific area, the user can also only collect remote sensing images of the specified area for training the photovoltaic land classification model, further reducing the amount of data and enabling the photovoltaic land classification model to obtain a better fitting effect in the specified area.

[0048] As a preferred implementation method of the embodiment of the present invention, this embodiment uses Sentinel-2 remote sensing images as data sources, selects 236 representative photovoltaic land areas from six continents, including Asia, Europe, Africa, North America, South America, and Oceania, which cover remote sensing images of different underlying land cover types and photovoltaic array arrangement methods in each continent, and these remote sensing images constitute a photovoltaic land sample data set. The constructed photovoltaic land sample data set is diverse and large enough to meet the training requirements of the photovoltaic land classification model and maximize the versatility of the model.

[0049] S2. Input the photovoltaic land sample data set into the constructed decision tree classifier for coarse classification to obtain the coarse segmentation result in the form of a binary raster map.

[0050] In the above step S2 of the present invention, in order to enhance the fitting effect of the photovoltaic land classification model on photovoltaic land in various surface environments, as much training data as possible is required. However, the existing photovoltaic land data is generally of low accuracy and insufficient coverage density on a global scale. If only the existing photovoltaic land data is used to train the photovoltaic land classification model, the fitting result will be insufficient in accuracy and versatility. Therefore, remote sensing images from different regions should be collected to construct fine samples distributed globally. Such images themselves not only have a huge data scale, but also the photovoltaic land portion in most images only accounts for a small part of the image. Therefore, these remote sensing images are roughly classified by a decision tree classifier, and the key parts with photovoltaic land are retained, so that the amount of data processing can be minimized while ensuring that the training samples are sufficiently rich.

[0051] It should be noted that in the present invention, the structure and classification method of the decision tree classifier belong to the prior art. It is a tree structure (can be a binary tree or a non-binary tree). Each non-leaf node represents a test on a feature attribute, each branch represents the output of this feature attribute on a certain value range, and each leaf node stores a category. The process of making decisions using a decision tree classifier is to start from the root node, test the corresponding feature attributes in the item to be classified, and select the output branch according to its value until the leaf node is reached, and the category stored in the leaf node is used as the decision result.

[0052] The classification limiting conditions of the decision tree classifier proposed in the present invention can be classified into two categories: photovoltaic land index conditions and other conditions. The photovoltaic land index conditions include the first photovoltaic land index, the second photovoltaic land index and the blue-red band ratio. The blue-red band ratio refers to the ratio of the blue band reflectivity to the red band reflectivity of the remote sensing image. The photovoltaic land index conditions are calculated for the purpose of highlighting the commonality of spectral characteristics of photovoltaic land in multiple bands in remote sensing images around the world. The calculation involves the reflectivity of five bands, namely the blue band, the red band, the near-infrared band, the first short-wave infrared band, and the second short-wave infrared band. The specific expression is:

[0053]

[0054]

[0055] Among them, ρ Blue ,ρ Red ,ρ NIR ,ρ SWIR1 ,ρ SWIR2 They represent the reflectivity of the blue band, red band, near-infrared band, first short-wave infrared band, and second short-wave infrared band respectively.

[0056] In this embodiment, the 458nm~523nm band range is used as the blue band, and its center wavelength is 490nm; the 650nm~680nm band range is used as the red band, and its center wavelength is 665nm; the 785nm~900nm band range is used as the near-infrared band, and its center wavelength is 842nm; the 1565nm~1655nm band range is used as the first short-wave infrared band, and its center wavelength is 1610nm; the 2100nm~2280nm band range is used as the second short-wave infrared band, and its center wavelength is 2190nm.

[0057] Other conditions of the decision tree classifier include slope, altitude, normalized building index (NDBI) and reflectivity of the first short-wave infrared band. The main purpose is to distinguish photovoltaic land from other objects with similarities in other images. By setting an upper limit for slope, photovoltaic land can be distinguished from mountain shadows; by setting an upper limit for altitude, misclassification of some high mountain areas can be eliminated; by the normalized building index (NDBI), photovoltaic land can be distinguished from some densely built-up areas; by the reflectivity of the first short-wave infrared band, photovoltaic land can be distinguished from dark water bodies.

[0058] Among the other conditions mentioned above, the calculation of slope and altitude requires elevation data, which can be obtained from the global digital elevation model (DEM). DEM is a digital simulation of the ground terrain through limited terrain elevation data. Currently, there are relatively mature global DEM data such as the Shuttle Radar Topography Mission (SRTM) and the Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model (ASTER GDEM). The Normalized Difference Building Index (NDBI) and the reflectivity of the first shortwave infrared band can be directly obtained from remote sensing images. Among them, the calculation formula of the Normalized Building Index is:

[0059]

[0060] As a preferred implementation method of an embodiment of the present invention, the basis for the decision tree coarse classification of the photovoltaic land sample data set is set as follows: simultaneously satisfying the first photovoltaic land index PVLI1∈[0.75,3.65], the second photovoltaic land index PVLI1∈[0.30,3.60], the blue-red band ratio BDRI∈[0.56,1.21], the slope SLOPE∈[0,20°], the altitude ALTITUDE∈[0,3500m], the normalized building index NDBI∈[-0.03,0.46] and the reflectivity ρ of the first short-wave infrared band SWIR1 ∈[0.07,0.58].

[0061] S3. Post-process the coarse segmentation results, set a buffer zone of a preset size around the post-processed coarse segmentation results, frame an outer envelope rectangle outside the buffer zone, and form a coarse classification area image of the photovoltaic land by the pixels inside the outer envelope rectangle in the remote sensing image.

[0062] It should be noted that in step S3, four post-processing techniques, namely spatial mean filtering, reclassification denoising, area threshold denoising and grid-vector conversion, are used to correct the rough segmentation result. The specific steps are as follows:

[0063] S31. First, perform spatial mean filtering on the coarse segmentation result.

[0064] In this embodiment, based on the preset domain range n, a sliding window of size (2n+1)×(2n+1) and all values ​​1 is used to perform mean convolution on the pixels in the rough segmentation result. The new value of the center pixel will be output as 1 only when all other pixels in the neighborhood specified by the center pixel are 1; otherwise, the output value of the center pixel will fall within the interval (0,1). The spatial mean filtering result retains the values ​​of the pixels at non-edge positions and weakens the values ​​of the pixels at broken edges or jagged edge protrusions.

[0065] S32. Then, the spatial mean filtering result is subjected to reclassification and denoising processing, that is, pixels with a value of 1 are retained, and pixels with a value of 1 are assigned a value of 0, so as to remove free small noise and smooth the edges of the coarsely classified photovoltaic land.

[0066] S33. Secondly, the reclassification and denoising result in the form of a binary raster image is vectorized to obtain a vectorized binary image.

[0067] S34. Finally, all elements in the vectorized binary image whose areas are larger than a preset area threshold are removed (i.e., the misclassified noise with a larger area that still exists after spatial mean filtering and classification denoising is removed), and a rough segmentation result after post-processing is obtained.

[0068] The purpose of post-processing the rough classification results is to make the edges of the rough segmentation results obtained by the decision tree classifier smoother and to repair some incompleteness caused by misclassification and omission. The final result is converted into vector data, which also makes the rough classification of photovoltaic land use results more independent and easy to edit.

[0069] It should also be noted that in the present invention, considering that the coarse classification method often cannot completely extract the entire photovoltaic land, or the recognition accuracy is insufficient, it is necessary to delineate a complete area with photovoltaic land around the coarse segmentation result, and input the corresponding remote sensing image into the constructed neural network model for training to obtain a model with better photovoltaic land recognition effect. In this embodiment, after post-processing the coarse segmentation results, through an overall investigation of the photovoltaic land sample data set, a size of 460m (slightly larger than ) buffer and framing an outer envelope rectangle outside the buffer can minimize missed points and ensure the integrity of the photovoltaic land in the input, which is in line with the goal of the present invention to maximize the use of key parts of photovoltaic land in the image with the minimum data processing amount.

[0070] As a preferred implementation of the embodiment of the present invention, the domain range of the spatial mean filter is set to 3 pixels, and the area threshold of the area threshold denoising is set to 2000m 2 , the buffer size is set to 460m. The final PV land rough classification area image is generated in this way, which has the best effect for the PV land sample data set from 236 sample areas as a whole.

[0071] S4. Generate a high-precision real label for each PV land coarse classification area image through human-computer interaction, and slice the PV land coarse classification area image and the real label of PV land. The sliced ​​PV land coarse classification area image and the real label of PV land constitute a PV land classification data set.

[0072] It should be noted that in step S4 of the present invention, the real labels of photovoltaic land can be generated with the help of tools such as Labelme and ArcGIS Pro. These tools currently provide open source access, and the generated real labels of photovoltaic land will automatically mark unique values ​​according to the type of land feature, which is convenient for training the photovoltaic land classification model. At the same time, in order to save workload, some photovoltaic land elements with better fitting effects in the photovoltaic land sample data set (GPVI) can be directly migrated to the photovoltaic land classification data set. After completing the production of the above-mentioned real labels, it is necessary to slice the coarse classification area image of photovoltaic land and the real label of photovoltaic land. The photovoltaic land classification model accepts square-shaped photovoltaic land coarse classification area image slices and photovoltaic land real label slices as input.

[0073] Furthermore, the photovoltaic land classification dataset can be divided into a training set and a test set in a ratio of 8:2 for training and testing the photovoltaic land classification model.

[0074] As a preferred implementation method of the present invention, by statistically analyzing the characteristics of photovoltaic land area in 236 sample areas around the world, it is determined that the average area of ​​a single photovoltaic map spot is about 0.07 km 2 Therefore, the size of the image slice of the photovoltaic land rough classification area is determined to be 64 pixels × 64 pixels (about 0.4 km 2 ), so that an average-sized PV land patch can occupy about 1 / 6 of the slice area, thus ensuring a certain degree of balance in the ratio of PV land to non-PV land in the slice.

[0075] S5. Train a photovoltaic land classification model that integrates multi-scale photovoltaic land features on the photovoltaic land classification dataset.

[0076] In the present invention, the photovoltaic land classification model is a main-sub-branch dual-branch structure, the main branch includes an input module, an encoder, a void space pyramid pooling module, a decoder and an output module, the sub-branch includes an input module, an encoder and five exponential space attention modules, wherein the encoder of the main branch, the encoder of the sub-branch and the decoder of the main branch each contain four residual blocks, the main branch is a U-shaped symmetrical structure, each residual block of the main branch decoder is jump-connected with each residual block of the same level of the main branch encoder, the input and output of each residual block of the sub-branch encoder are densely connected, the exponential feature space weight generated by each exponential space attention module in the sub-branch is densely connected with each residual block of the same level of the main branch encoder, the exponential feature space weight generated by each exponential space attention module in the sub-branch is densely connected with the exponential feature space weight generated by the next exponential space attention module, and then densely connected with each residual block of the same level of the main branch decoder.

[0077] It should be noted that in the photovoltaic land classification model in step S5, Figure 3 As shown, the specific processing flow is as follows: based on the rough classification area image of the photovoltaic land after slicing, a multispectral image and an index image are generated, the multispectral image is input into the input module of the main branch to obtain a first spectral feature map, the index image is input into the input module of the secondary branch to obtain a first index feature map, the first index feature map is input into the first index space attention module of the secondary branch to obtain a first index feature space weight matrix, and the first index feature space weight matrix and the first spectral feature map are multiplied to obtain a first fusion feature map;

[0078] Input the first fused feature map into the first residual block of the main branch encoder to obtain a second spectral feature map, input the first exponential feature map into the first residual block of the sub-branch encoder to obtain a second exponential feature map, densely connect the first exponential feature map and the second exponential feature map to obtain a first densely connected feature map, input the first densely connected feature map into the second exponential space attention module of the sub-branch to obtain a second exponential feature space weight matrix, and multiply the second exponential feature space weight matrix and the second spectral feature map to obtain a second fused feature map;

[0079] Input the second fused feature map into the second residual block of the main branch encoder to obtain a third spectral feature map, input the first densely connected feature map into the second residual block of the sub-branch encoder to obtain a third exponential feature map, densely connect the third exponential feature map and the first densely connected feature map to obtain a second densely connected feature map, input the second densely connected feature map into the third exponential space attention module of the sub-branch to obtain a third exponential feature space weight matrix, and multiply the third exponential feature space weight matrix and the third spectral feature map to obtain a third fused feature map;

[0080] Input the third fused feature map into the third residual block of the main branch encoder to obtain a fourth spectral feature map, input the second densely connected feature map into the third residual block of the sub-branch encoder to obtain a fourth exponential feature map, densely connect the fourth exponential feature map and the second densely connected feature map to obtain a third densely connected feature map, input the third densely connected feature map into the fourth exponential space attention module of the sub-branch to obtain a fourth exponential feature space weight matrix, and multiply the fourth exponential feature space weight matrix and the fourth spectral feature map to obtain a fourth fused feature map;

[0081] Input the fourth fused feature map into the fourth residual block of the main branch encoder to obtain a fifth spectral feature map, input the third densely connected feature map into the fourth residual block of the sub-branch encoder to obtain a fifth exponential feature map, densely connect the fifth exponential feature map and the third densely connected feature map to obtain a fourth densely connected feature map, input the fourth densely connected feature map into the fifth exponential space attention module of the sub-branch to obtain a fifth exponential feature space weight matrix, and multiply the fifth exponential feature space weight matrix and the fifth spectral feature map to obtain a fifth fused feature map;

[0082] Input the fifth fused feature map into the atrous space pyramid pooling module to obtain the atrous space pooling feature map, multiply the atrous space pooling feature map and the fifth exponential feature space weight matrix to obtain the sixth fused feature map, input the sixth fused feature map into the fourth residual block of the main branch decoder to obtain the sixth spectral feature map, densely connect the fifth exponential feature space weight matrix and the fourth exponential feature space weight matrix to obtain the first dense connection weight matrix, and multiply the sixth spectral feature map and the first dense connection weight matrix to obtain the seventh fused feature map;

[0083] Input the seventh fused feature map into the third residual block of the main branch decoder to obtain a seventh spectral feature map, densely connect the first dense connection weight matrix and the third exponential feature space weight matrix to obtain a second dense connection weight matrix, and multiply the seventh spectral feature map and the second dense connection weight matrix to obtain an eighth fused feature map;

[0084] Inputting the eighth fused feature map into the second residual block of the main branch decoder to obtain an eighth spectral feature map, densely connecting the second dense connection weight matrix and the second exponential feature space weight matrix to obtain a third dense connection weight matrix, and multiplying the eighth spectral feature map and the third dense connection weight matrix to obtain a ninth fused feature map;

[0085] Input the ninth fused feature map into the first residual block of the main branch decoder to obtain a ninth spectral feature map, densely connect the third dense connection weight matrix and the first exponential feature space weight matrix to obtain a fourth dense connection weight matrix, and multiply the ninth spectral feature map and the fourth dense connection weight matrix to obtain a tenth fused feature map;

[0086] Finally, the tenth fused feature map is input into the output module of the main branch to obtain the segmentation result output by the photovoltaic land classification model.

[0087] It should be noted that the present invention aims at the goal of accurately identifying photovoltaic land in remote sensing images, taking into account the multi-band characteristics of remote sensing images that are more complex than natural images and the index features derived from spectral band features based on domain knowledge (and such features have been proven to help improve segmentation accuracy), and constructs a photovoltaic land classification model (PVLI-FSSM). The ResUNet++ model is used as the baseline model, and its structure is improved accordingly according to the rich spectral information and index features of remote sensing images, which are specifically manifested as follows:

[0088] 1) Dual-branch structure: The photovoltaic land classification model integrates the photovoltaic land index features into the spectral information of the main branch at multiple scales through the sub-branch, the index spatial attention module, and multiple connection methods between the main and sub-branch modules. The main branch is a ResUNet++ structure, which is responsible for extracting the multi-scale and contextual features of the spectral channel of the input image, and integrating the index features in the sub-branch into the spectral information through multiple connection methods. The main branch includes an input module, an encoder, an atrous spatial pyramid pooling module (ASPP), a decoder, and an output module. The sub-branch includes an input module, an encoder, and five index spatial attention modules (IFSA). The encoder of the main branch, the encoder of the sub-branch, and the decoder of the main branch each contain four residual blocks (ResBlock). Each residual block of the main branch decoder is jump-connected with each residual block of the same level in the main branch encoder. The output module is responsible for converting the output of the decoder into a segmentation result.

[0089] 2) Special connection methods: In order to maximize the use of the index and the multi-scale integration with spectral characteristics, in addition to the ordinary connection, four special connection methods are designed in the photovoltaic land classification model:

[0090] 2.1) Dense connections are made between the input and output of each residual block of the secondary branch encoder;

[0091] 2.2) The exponential feature space weights generated by each exponential spatial attention module in the secondary branch are densely connected with the residual blocks ResBlock at the same level of the main branch encoder;

[0092] 2.3) The exponential feature space weight matrix (IFSW) generated by each exponential space attention module in the secondary branch is densely connected with each residual block at the same level of the main branch decoder;

[0093] 2.4) Before the exponential spatial attention module in the secondary branch is densely connected to the main branch decoder, the exponential feature space weight matrix generated by each exponential spatial attention module in the secondary branch is densely connected with the exponential feature space weight matrix generated by the next exponential spatial attention module.

[0094] Among them, the third and fourth connections are designed to highlight the foreground, suppress the background, and refine the boundaries.

[0095] 3) Atrous Spatial Pyramid Pooling Module (ASPP): In the connection part between the main branch encoder and decoder, the photovoltaic land classification model uses the atrous spatial pyramid pooling module to replace the connection layer of ResUNet.

[0096] 4) Index Spatial Attention Module IFSA: The purpose of the spatial attention mechanism is to improve the recognition ability of specific targets in remote sensing images. It can highlight the multi-class features of the target by multiplying the multi-class features of the target. In the neural network, spatial attention has the function of capturing and highlighting features. Through a variety of statistical modules or convolution combination modules, the dimension is reduced and the spatial weight of a certain type of feature is calculated to achieve the spatial position importance analysis of the feature. The output feature spatial weight is operated with other types of features to enhance the specific target features. This process can also be understood as a fusion.

[0097] The photovoltaic land classification model proposed in the present invention introduces an exponential spatial attention module based on spatial attention in the secondary branch. The input of the first exponential spatial attention module is the index image of the C layer. After passing through the first exponential spatial attention module, the first exponential feature space weight matrix IFSW is output and input into the encoder and decoder of the main branch, and fused with the spectral feature maps of each scale output in the main branch to highlight the characteristics of the photovoltaic land.

[0098] In the exponential space attention module, its input continues to mine features through branches such as the ASPP module, the Max module (calculating the maximum value across channels), and the Mean module (calculating the mean across channels). The three branches reduce the dimension of the input exponential feature (IF) from N×C×H×W to N×1×H×W, and then concatenate the outputs of the three branches to obtain a feature map of N×3×H×W, which is then added to the input feature map in a residual connection manner, and processed into a layer 1 with a value range of [0,1] through a 3×3 convolution layer and a Sigmoid activation function, i.e., the exponential feature space weight matrix.

[0099] 5) Multi-channel input: The multispectral image channels input into the main branch of the photovoltaic land classification model are the blue band, green band, red band and the first short-wave infrared band. Here, the green band corresponds to the band range of 543nm~578nm, and its central wavelength is 560nm; the index image channels input into the sub-branch are the first photovoltaic land index and the second photovoltaic land index. The bands involved in the index calculation are: blue band, red band, near-infrared band, first short-wave infrared band and second short-wave infrared band. The calculation of all indexes is completed before inputting the sub-branch.

[0100] The above multispectral image and index image are obtained in the following way:

[0101] A) In the rough classification area image of the photovoltaic land after slicing, the images of the blue band, green band, red band and the first short-wave infrared band are taken for band synthesis to form a multispectral image.

[0102] B) In the image of the coarse classification area of ​​photovoltaic land after slicing, the images of the red band, the near-infrared band, the first short-wave infrared band and the second short-wave infrared band are taken and calculated according to the calculation formula of the first photovoltaic land index to obtain the first photovoltaic land index image; in the image of the coarse classification area of ​​photovoltaic land after slicing, the images of the blue band, the red band, the near-infrared band, the first short-wave infrared band and the second short-wave infrared band are taken and calculated according to the calculation formula of the second photovoltaic land index to obtain the second photovoltaic land index image; the first photovoltaic land index image and the second photovoltaic land index image are band synthesized to form an index image.

[0103] In order to help those skilled in the art better understand the structure of the photovoltaic land classification model, the principle of the ResUNet++ model is now explained as follows.

[0104] The ResUNet++ model is a variant of the classic convolutional neural network (CNN) - UNet, which is a deep learning model with a simple structure, clear principles and high segmentation accuracy, and is often used in the field of remote sensing image interpretation and recognition. UNet consists of three parts: encoder, connection layer and decoder. The encoder and decoder both contain four convolution modules, and each module contains two consecutive convolution layers, but the sampling method after the convolution layer of the encoder and decoder modules is different. After each convolution module of the encoder, the maximum pooling layer Max Pooling is used to achieve downsampling, and the dimension is halved to achieve the purpose of extracting more abstract features layer by layer; before each convolution module of the decoder is an upsampling layer, so that the number of features in each layer is halved compared to the next layer (backward layer), so as to achieve the purpose of restoring the spatial size layer by layer. The advantages of UNet are: each layer of the encoder is copied, cropped and connected to the corresponding upsampling layer in the decoder, which retains the shallow information and provides more detailed information for the decoder when upsampling, thus avoiding the problem of reducing shallow features due to increased depth.

[0105] ResUNet++ replaces the convolution module in the original UNet with the residual module ResBlock, and uses Skip Connection to achieve cross-layer connection in each residual block, so that the model has the advantages of both the residual connection of ResNet and the context information fusion of UNet.

[0106] Furthermore, the ResUNet++ model makes the following improvements to the UNet network architecture:

[0107] 1) Compression and Excitation (SE) module,The encoder part of the network contains four downsampling modules. Except for the first downsampling module, the remaining downsampling modules contain a residual module and a SE module (i.e., an SE module is added after each residual module).

[0108] 2) Attention module. The decoder part of the network contains four upsampling modules. Except for the last downsampling module, the remaining upsampling modules contain an Attention module, an upsampling layer and a residual module, that is, an Attention module is added before each upsampling layer.

[0109] It should be noted that in the present invention, the photovoltaic land classification model PVLI-FSSM needs to use multiple preset hyperparameters during training, namely, training round E, batch size B, optimizer O, initial learning rate Lr, momentum coefficient M, and weight decay W.

[0110] As a preferred implementation method of an embodiment of the present invention, the hyperparameters when training the photovoltaic land classification model are set as follows: the training rounds E are set to 300 rounds, the batch size B is set to 4, the optimizer O is set to SGD, the initial learning rate Lr is set to 0.01, the momentum coefficient M is set to 0.9, and the weight decay W is set to 1e-4.

[0111] S6. Input the image of the coarse classification area of ​​the photovoltaic land to be classified into the trained photovoltaic land classification model to obtain the segmentation result of the image of the coarse classification area of ​​the photovoltaic land, so as to realize the fine identification of the photovoltaic land.

[0112] It should be noted that, after experiments, although images from other sensors can also successfully identify photovoltaic land after being classified by the photovoltaic land classification model, the final recognition effect is often better when the image types of the training set and the validation set are consistent. At the same time, the method of the present invention requires that the image has sufficient spectral information, so the aforementioned methods are all based on optical remote sensing images. When selecting any image for photovoltaic land identification, it is necessary to ensure that its wavelength range is within the range of 380nm to 1648nm.

[0113] It should be noted that when processing the identification of photovoltaic land in a large area, the scale of image data input by the user into the photovoltaic land classification model may be large. At this time, the key parts containing photovoltaic land in the remote sensing image can be preliminarily extracted through the aforementioned S2-S3 coarse classification and post-processing process, and then the key parts are input into the photovoltaic land classification model to generate identification results, and then the identification results are imported into the image of the entire area. This can reduce the amount of data processing, improve the calculation speed, and obtain reliable and high-precision photovoltaic land identification results.

[0114] It should be noted that in the above step S6, the segmentation result generated by the photovoltaic land classification model already has a high degree of accuracy, but it can also be slightly modified according to user needs. After removing a small amount of misclassification and omissions, the corrected segmentation result is converted from raster to vector and output as polygon photovoltaic land product-level data. The purpose of converting the segmentation result from raster data to vector data here is to facilitate the numerical statistics and editing operations such as area by partition, and the user can also choose the appropriate data output format according to actual needs.

[0115] In summary, if Figure 2 As shown, the present invention proposes a "two-step" joint decision tree and deep learning large-area photovoltaic land classification method, which specifically includes:

[0116] Step 1: Delineation of the rough classification area of ​​photovoltaic land based on the decision tree classifier. Based on the GEE remote sensing cloud platform and Sentinel-2 remote sensing images, the constructed decision tree classifier is used to obtain the rough segmentation results in the form of binary raster images. After spatial mean filtering, reclassification denoising, area threshold denoising and raster-vector conversion, the rough classification area image of photovoltaic land is formed.

[0117] Step 2: Accurate segmentation of photovoltaic land in the coarse classification area based on deep learning. Based on the trained photovoltaic land classification model, photovoltaic land is accurately extracted from the Sentinel-2 image slices in the coarse classification area, and after manual verification, data correction, and raster-vector conversion, product-level photovoltaic land vector data is obtained.

[0118] The present invention will further demonstrate the detailed implementation process of the large-area photovoltaic land classification method based on the joint decision tree and deep learning shown in S1 to S6 in a specific example below, so as to facilitate the understanding of the essence of the present invention.

[0119] Example

[0120] In this embodiment, the Sentinel-2 remote sensing satellite image is used as the source data, and the specific implementation process of the large-area photovoltaic land classification method using a joint decision tree and deep learning is as described above and will not be repeated here.

[0121] In this embodiment, remote sensing images of representative areas are first selected according to the global photovoltaic land vector dataset (GPVI). In this embodiment, regional images of photovoltaic land construction from 236 representative areas in six continents, including Asia, Europe, Africa, North America, South America, and Oceania, are selected to form a global representative photovoltaic land remote sensing image dataset. Since the cloud and fog occlusion of remote sensing data will affect the identification of image objects, it is necessary to first pre-process the global representative photovoltaic land remote sensing image dataset, remove images with cloud cover greater than 20% when acquiring data, and perform cloud removal operations. After that, the spectral index and slope, etc. are calculated for each image, and input into the decision tree classifier, and the photovoltaic land and non-photovoltaic land are distinguished according to the pre-set limiting conditions to form a binary raster map. Then, the result is subjected to four post-processing steps of spatial mean filtering, reclassification denoising, area threshold denoising, and vectorization to make it smoother, and finally a buffer zone and an outer envelope rectangle are generated around the result. Based on the outer envelope rectangle, a small number of key areas with photovoltaic land can be extracted from each image through interception operations to form a coarse classification image block. In this embodiment, the above steps are all completed on the Google Earth Engine online cloud platform (code.earthengine.com). The decision tree is divided into photovoltaic land use restrictions: PVLI1>1.142∩PVLI2>1.11∩BDRI>0.674. At the same time, it is also necessary to meet the slope less than 20°, the altitude is less than 3500m, the normalized building index (NDBI) is greater than 0.075, and the reflectivity of the 1572nm~1648nm band is greater than 0.194 (mainly used to distinguish water bodies). The elevation data required to calculate the slope and altitude are directly obtained by the Google Earth Engine platform based on the global digital elevation model (DEM) provided by the Space Shuttle Radar Topography Mission (SRTM). In the post-processing technology, the domain range of the spatial mean filter is set to 3 pixels, and the area threshold set for area threshold denoising is 2000m 2 , the surrounding buffer size is set to 460m.

[0122] After obtaining the image blocks to be classified in the above key areas, in the ArcGIS Pro tool, human-computer interaction is performed by creating features to generate high-precision photovoltaic land labels, and then the images and labels are sliced ​​into 64 pixels × 64 pixels according to the requirements of the PVLI-FSSM model. If the size of some boundary images does not meet the slicing requirements, the surrounding area with a value of 0 is filled. Slicing finally obtained 10,219 pairs of images and labels, which were divided into training sets and test sets in a ratio of 8:2. The training set containing images and corresponding labels was input into the PVLI-FSSM model for iterative training. After training, the hyperparameters were adjusted according to the performance of the test set. Finally, the optimal preset hyperparameters were determined: 300 rounds of model training, a batch size of 4, an optimizer of SGD, an initial learning rate of 0.01, a momentum coefficient of 0.9, and a weight decay of 1e-4. After the PVLI-FSSM model is trained, it can accurately identify the photovoltaic land parts in remote sensing images around the world.

[0123] In this embodiment, in order to verify the accuracy of the PVLI-FSSM model combined with the spectral index features obtained by the above training, its performance was evaluated by the overall accuracy (OA) and mean intersection over union (mIoU) indicators. The results showed that OA = 97.54%, mIoU = 95.02%, indicating that the model has good fitting ability.

[0124] It can be seen that after training the photovoltaic land classification model, large-scale photovoltaic land identification can be carried out for any area that needs to be identified. After raster-to-vector conversion operations, the final photovoltaic land feature class data with precise details and complete edges can be obtained.

[0125] In this embodiment, large-area photovoltaic land mapping is performed with Jiangsu Province and Ningxia Hui Autonomous Region as representatives. The results are shown in Tables 1 and 2. The overall classification accuracy of the mapping results reaches 95.56%, and the deep learning data processing volume is reduced to 4.5% of the original, achieving rapid and accurate mapping of large-area photovoltaic land. Accurate photovoltaic land product data can be obtained based on provincial-level remote sensing images of different years, and the temporal and spatial changes of photovoltaic land in the two provinces can be understood based on inter-annual data.

[0126] Table 1. Photovoltaic land identification accuracy table of 6 typical inspection areas in Jiangsu

[0127]

[0128] Table 2. Photovoltaic land identification accuracy table of six typical inspection areas in Ningxia

[0129]

[0130] At the same time, in order to compare the effect of multi-scale fusion of spectral index features on the accuracy of the PVLI-FSSM model in the present invention, an ablation experiment was carried out on the photovoltaic land dataset, that is, the PVLI-FSSM model in the present invention was compared with the existing baseline ResUNet++ model. After 100 rounds of training, the two models were compared in segmentation effect, and the results are shown in Table 3. The results show that the PVLI-FSSM model combined with spectral index features can better identify the photovoltaic land part in the remote sensing image, and the introduction of spectral index features has a significant improvement in accuracy, indicating that the spectral index features based on multi-scale fusion can further distinguish the photovoltaic land from the image background, thereby achieving accurate identification.

[0131] Table 3. Accuracy comparison between the baseline ResUNet++ model and the PVLI_FSSM model

[0132]

[0133] In summary, the present invention is based on the rapid development of geographic remote sensing big data and deep learning algorithms. Aiming at the problem of large data scale and lack of globally distributed training samples in large-area photovoltaic land identification research, a large-area photovoltaic land classification method combining decision trees and deep learning is provided, and a representative global photovoltaic deep learning data set is constructed. At the same time, the deep learning method is used to give full play to its strong characterization ability of nonlinear relationships and its ability to explore deep features, and long-term high-precision photovoltaic land product data represented by Jiangsu and Ningxia are constructed, proving that the present invention can serve to improve the understanding of the spatiotemporal differentiation characteristics and change trends of global photovoltaic land, and provide a way to analyze the construction of global energy infrastructure.

[0134] It should also be noted that the large-area photovoltaic land classification method of the combined decision tree and deep learning in the above embodiment can essentially be executed by a computer program or module. Therefore, similarly, based on the same inventive concept, another preferred embodiment of the present invention also provides a combined decision tree and deep learning large-area photovoltaic land classification system corresponding to the combined decision tree and deep learning large-area photovoltaic land classification method provided in the above embodiment, such as Figure 4 As shown, it includes:

[0135] The data acquisition module is used to select data of representative photovoltaic land areas from the global photovoltaic land location data set and classification data set, and construct a photovoltaic land sample data set containing remote sensing images, geographic coordinates, imaging time, underlying land cover type and photovoltaic array arrangement mode;

[0136] A coarse classification module is used to input the photovoltaic land sample data set into the constructed decision tree classifier for coarse classification, and obtain a coarse segmentation result in the form of a binary raster map;

[0137] A post-processing module is used to post-process the rough segmentation results, set a buffer zone of a preset size around the post-processed rough segmentation results, frame an outer envelope rectangle outside the buffer zone, and form a rough classification area image of photovoltaic land from the pixels inside the outer envelope rectangle in the remote sensing image;

[0138] The dataset construction module is used to generate a real photovoltaic land label for each photovoltaic land coarse classification area image through human-computer interaction, and to slice the photovoltaic land coarse classification area image and the photovoltaic land real label, so that the photovoltaic land classification dataset is formed by the sliced ​​photovoltaic land coarse classification area image and the photovoltaic land real label;

[0139] Model training module, used to train a photovoltaic land classification model that integrates photovoltaic land features at multiple scales on the photovoltaic land classification dataset;

[0140] The result acquisition module is used to input the coarse classification area image of the photovoltaic land to be classified into the trained photovoltaic land classification model, obtain the segmentation result of the coarse classification area image of the photovoltaic land, and realize the fine identification of the photovoltaic land.

[0141] It should also be noted that those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the system described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here. In the various embodiments provided in this application, the division of steps or modules in the system and method is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules or steps can be combined or integrated together, and a module or step can also be split.

[0142] The above-described embodiment is only a preferred solution of the present invention, but it is not intended to limit the present invention. A person skilled in the relevant technical field may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, any technical solution obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present invention.

Claims

1. A large-area photovoltaic land classification method combining decision tree and deep learning, characterized in that: The following steps are involved: S1. Select data of representative photovoltaic land areas from the global photovoltaic land location dataset and classification dataset, and construct a photovoltaic land sample dataset containing remote sensing images, geographic coordinates, imaging time, underlying land cover type, and photovoltaic array arrangement mode; S2, inputting the photovoltaic land sample data set into the constructed decision tree classifier for rough classification, and obtaining a rough segmentation result in the form of a binary raster map; S3, post-processing the coarse segmentation result, setting a buffer zone of a preset size around the post-processed coarse segmentation result, framing an outer envelope rectangle outside the buffer zone, and forming a coarse classification area image of the photovoltaic land by the pixels inside the outer envelope rectangle in the remote sensing image; S4. Generate a real label for each photovoltaic land rough classification area image through human-computer interaction, and slice the photovoltaic land rough classification area image and the real label of photovoltaic land, so that the photovoltaic land classification data set is formed by the sliced ​​photovoltaic land rough classification area image and the real label of photovoltaic land; S5. Train a photovoltaic land classification model that integrates multi-scale photovoltaic land features on the photovoltaic land classification dataset; S6. Input the image of the coarse classification area of ​​the photovoltaic land to be classified into the trained photovoltaic land classification model to obtain the segmentation result of the image of the coarse classification area of ​​the photovoltaic land, so as to realize the fine identification of the photovoltaic land.

2. A large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 1, characterized in that: In step S2, the classification limiting conditions of the decision tree classifier are divided into photovoltaic land index conditions and other conditions. The photovoltaic land index conditions are the first photovoltaic land index, the second photovoltaic land index and the blue-red band ratio. The other conditions are slope, altitude, normalized building index and reflectivity of the first short-wave infrared band.

3. A large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 2, characterized in that: The specific calculation method of the first photovoltaic land index PVLI1, the second photovoltaic land index PVLI2, the blue-red band ratio BDRI, and the normalized building index NDBI is as follows: Among them, ρ Blue ,ρ Red ,ρ NIR ,ρ SWIR1 ,ρ SWIR2 They represent the reflectivity of the blue band, red band, near-infrared band, first short-wave infrared band, and second short-wave infrared band respectively.

4. A large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 2 or 3, characterized in that: In step S2, the basis for the rough classification of the photovoltaic land sample data set is set as follows: the first photovoltaic land index PVLI1∈[0.75,3.65], the second photovoltaic land index PVLI1∈[0.30,3.60], the blue-red band ratio BDRI∈[0.56,1.21], the slope SLOPE∈[0,20°], the altitude ALTITUDE∈[0,3500m], the normalized building index NDBI∈[-0.03,0.46] and the reflectivity ρ of the first short-wave infrared band are simultaneously satisfied. SWIR1 ∈[0.07,0.58].

5. The large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 1, characterized in that: In step S3, the specific process of post-processing the coarse segmentation result is: first, the coarse segmentation result is subjected to spatial mean filtering, then the spatial mean filtering result is subjected to reclassification and denoising, and then the reclassification and denoising result in the form of a binary raster image is vectorized to obtain a vectorized binary image, and finally, all elements in the vectorized binary image whose area is greater than a preset area threshold are removed to obtain a post-processed coarse segmentation result.

6. A large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 3, characterized in that: The photovoltaic land classification model is a main-sub dual-branch structure, wherein the main branch includes an input module, an encoder, a void space pyramid pooling module, a decoder and an output module, and the sub-branch includes an input module, an encoder and five exponential space attention modules, wherein the encoder of the main branch, the encoder of the sub-branch and the decoder of the main branch each include four residual blocks, the main branch is a U-shaped symmetrical structure, each residual block of the main branch decoder is jump-connected with each residual block of the same level of the main branch encoder, and the input and output of each residual block of the sub-branch encoder are densely connected, the exponential feature space weight generated by each exponential space attention module in the sub-branch is densely connected with each residual block of the same level of the main branch encoder, and the exponential feature space weight generated by each exponential space attention module in the sub-branch is densely connected with the exponential feature space weight generated by the next exponential space attention module, and then densely connected with each residual block of the same level of the main branch decoder.

7. A large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 6, characterized in that: In the photovoltaic land classification model of step S5, the specific processing flow is: based on the rough classification area image of the photovoltaic land after slicing, a multispectral image and an index image are generated, the multispectral image is input into the input module of the main branch to obtain a first spectral feature map, the index image is input into the input module of the sub-branch to obtain a first index feature map, the first index feature map is input into the first index space attention module of the sub-branch to obtain a first index feature space weight matrix, the first index feature space weight matrix and the first spectral feature map are multiplied to obtain a first fused feature map; the first fused feature map is input into the first residual block of the main branch encoder to obtain a second spectral feature map, and the first index feature map is input into the sub-branch. In the first residual block of the encoder, a second exponential feature map is obtained, the first exponential feature map and the second exponential feature map are densely connected to obtain a first densely connected feature map, the first densely connected feature map is input into the second exponential space attention module of the secondary branch to obtain a second exponential feature space weight matrix, the second exponential feature space weight matrix is ​​multiplied with the second spectral feature map to obtain a second fused feature map; the second fused feature map is input into the second residual block of the main branch encoder to obtain a third spectral feature map, the first densely connected feature map is input into the second residual block of the secondary branch encoder to obtain a third exponential feature map, the third exponential feature map and the first densely connected feature map are densely connected, A second densely connected feature map is obtained, and the second densely connected feature map is input into the third exponential space attention module of the sub-branch to obtain a third exponential feature space weight matrix, and the third exponential feature space weight matrix and the third spectral feature map are multiplied to obtain a third fused feature map; the third fused feature map is input into the third residual block of the main branch encoder to obtain a fourth spectral feature map, and the second densely connected feature map is input into the third residual block of the sub-branch encoder to obtain a fourth exponential feature map, and the fourth exponential feature map and the second densely connected feature map are densely connected to obtain a third densely connected feature map, and the third densely connected feature map is input into the fourth exponential space attention module of the sub-branch to obtain the four exponential feature space weight matrices, multiplying the fourth exponential feature space weight matrix and the fourth spectral feature map to obtain a fourth fused feature map; inputting the fourth fused feature map into the fourth residual block of the main branch encoder to obtain a fifth spectral feature map, inputting the third densely connected feature map into the fourth residual block of the sub-branch encoder to obtain a fifth exponential feature map, densely connecting the fifth exponential feature map and the third densely connected feature map to obtain a fourth densely connected feature map, inputting the fourth densely connected feature map into the fifth exponential space attention module of the sub-branch to obtain a fifth exponential feature space weight matrix, and multiplying the fifth exponential feature space weight matrix and the fifth spectral feature map to obtain a fifth fused feature map;Input the fifth fused feature map into the atrous space pyramid pooling module to obtain the atrous space pooling feature map, multiply the atrous space pooling feature map and the fifth exponential feature space weight matrix to obtain the sixth fused feature map, input the sixth fused feature map into the fourth residual block of the main branch decoder to obtain the sixth spectral feature map, densely connect the fifth exponential feature space weight matrix and the fourth exponential feature space weight matrix to obtain the first dense connection weight matrix, and multiply the sixth spectral feature map and the first dense connection weight matrix to obtain the seventh fused feature map; Input the seventh fused feature map into the third residual block of the main branch decoder to obtain the seventh spectral feature map, densely connect the first dense connection weight matrix and the third exponential feature space weight matrix to obtain the second dense connection weight matrix, and multiply the seventh spectral feature map and the second dense connection weight matrix to obtain the eighth fused feature map; input the eighth fused feature map into the second residual block of the main branch decoder to obtain the eighth spectral feature map, densely connect the second dense connection weight matrix and the second exponential feature space weight matrix to obtain the third dense connection weight matrix, and multiply the eighth spectral feature map and the third dense connection weight matrix to obtain the ninth fused feature map; input the ninth fused feature map into the first residual block of the main branch decoder to obtain the ninth spectral feature map, densely connect the third dense connection weight matrix and the first exponential feature space weight matrix to obtain the fourth dense connection weight matrix, and multiply the ninth spectral feature map and the fourth dense connection weight matrix to obtain the tenth fused feature map; The tenth fused feature map is input into the output module of the main branch to obtain the segmentation result output by the photovoltaic land classification model.

8. A large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 7, characterized in that: The multispectral image is obtained in the following manner: In the image of the coarsely classified area of ​​photovoltaic land after slicing, the images of the blue band, green band, red band and the first short-wave infrared band are taken for band synthesis to form a multispectral image.

9. A large-area photovoltaic land classification method combining decision tree and deep learning as claimed in claim 7, characterized in that: The method for obtaining the index image is specifically as follows: The images of the red band, near-infrared band, first short-wave infrared band and second short-wave infrared band are taken from the image of the coarse classification area of ​​photovoltaic land after slicing, and are calculated according to the calculation formula of the first photovoltaic land index to obtain the first photovoltaic land index image; the images of the blue band, red band, near-infrared band, first short-wave infrared band and second short-wave infrared band are taken from the image of the coarse classification area of ​​photovoltaic land after slicing, and are calculated according to the calculation formula of the second photovoltaic land index to obtain the second photovoltaic land index image; the first photovoltaic land index image and the second photovoltaic land index image are band synthesized to form an index image.

10. A large-area photovoltaic land classification system combining decision trees and deep learning, characterized in that: include: The data acquisition module is used to select data of representative photovoltaic land areas from the global photovoltaic land location data set and classification data set, and construct a photovoltaic land sample data set containing remote sensing images, geographic coordinates, imaging time, underlying land cover type, and photovoltaic array arrangement mode; A coarse classification module is used to input the photovoltaic land sample data set into the constructed decision tree classifier for coarse classification, and obtain a coarse segmentation result in the form of a binary raster map; A post-processing module is used to post-process the rough segmentation results, set a buffer zone of a preset size around the post-processed rough segmentation results, frame an outer envelope rectangle outside the buffer zone, and form a rough classification area image of photovoltaic land from the pixels inside the outer envelope rectangle in the remote sensing image; The dataset construction module is used to generate a real photovoltaic land label for each photovoltaic land coarse classification area image through human-computer interaction, and to slice the photovoltaic land coarse classification area image and the photovoltaic land real label, so that the photovoltaic land classification dataset is formed by the sliced ​​photovoltaic land coarse classification area image and the photovoltaic land real label; Model training module, used to train a photovoltaic land classification model that integrates photovoltaic land features at multiple scales on the photovoltaic land classification dataset; The result acquisition module is used to input the coarse classification area image of the photovoltaic land to be classified into the trained photovoltaic land classification model, obtain the segmentation result of the coarse classification area image of the photovoltaic land, and realize the fine identification of the photovoltaic land.

Images