Direct irradiance hybrid prediction method based on ground-based cloud picture multi-feature extraction

Abstract

The invention relates to a direct irradiance hybrid prediction model based on multi-feature extraction of a ground-based cloud picture, which combines knowledge-based cloud detection and an improved deep convolutional neural network, extracts global features and local features as input of a feature fusion and irradiance prediction module, and uses the extracted global features and local features as input data together with a solar elevation angle. The features are comprehensively processed through a multi-layer full-connection network, so that the complex relation between the features is learned; and finally, giving a predicted value of the direct irradiance through a regression output layer. According to the method, the problem of direct solar radiation prediction under complex meteorological conditions is fully considered and dealt with, and the model can describe cloud layer features more comprehensively through joint extraction of global and local features, so that high-precision irradiance prediction is realized; experimental results verify the effectiveness and superiority of the method, and particularly, the method shows relatively high prediction precision and stability under complex and variable weather conditions.

Description

Technical Field

[0001] The invention belongs to the technical field of new energy prediction, and specifically relates to a direct irradiance hybrid prediction method based on multi-feature extraction of ground-based cloud images. Background Art

[0002] In recent years, with the increasing global dependence on fossil energy and the increasingly prominent environmental pollution problem, solar energy as a sustainable and clean energy has gradually attracted widespread attention. Photovoltaic power generation is considered to be one of the important means to achieve carbon neutrality due to its environmental protection and abundant resources. However, due to complex meteorological conditions, especially changes in cloud cover, the direct normal irradiance (DNI) of the sun fluctuates significantly, which affects the power generation stability and efficiency of the photovoltaic system. Therefore, how to effectively predict the direct normal irradiance of the sun with high precision has become an important link in the prediction and scheduling of photovoltaic power generation systems.

[0003] At present, the mainstream direct solar irradiance (DNI) prediction methods can be divided into knowledge-based cloud detection methods and deep learning-based feature extraction methods. Knowledge-based cloud detection methods mainly rely on manually extracted cloud features, such as cloud coverage and transparency. These methods perform well when weather conditions are relatively simple, but they show obvious shortcomings when dealing with changeable and complex meteorological conditions and lack the ability to adapt to the diversity and randomness of clouds. Deep learning methods, such as convolutional neural networks (CNNs), have excellent performance in image recognition and prediction due to their powerful feature extraction and nonlinear fitting capabilities. However, due to the complexity of cloud changes, the local features extracted by deep learning methods alone are often not enough for the CNN model to accurately describe the overall weather conditions, especially when capturing long-span and large-scale cloud information, there are certain limitations.

[0004] Cloud variations can cause fluctuations in direct solar irradiance (DNI), which in turn leads to significant changes in photovoltaic (PV) power generation. Therefore, DNI prediction based on cloud characteristics is an important support for achieving accurate PV power forecasts. However, modeling the relationship between DNI and cloud conditions remains challenging due to complex cloud variations and ambiguity of key features.

[0005] Therefore, the purpose of this invention is to propose a novel hybrid model that integrates knowledge-based cloud detection with an improved convolutional neural network, which can simultaneously extract global and local features of clouds, thereby improving the prediction accuracy of direct solar radiation under complex weather conditions. Summary of the invention

[0006] In view of the problems existing in the prior art, the present invention proposes a hybrid prediction model of direct irradiance based on multi-feature extraction of ground-based cloud images. By integrating knowledge-based cloud detection technology and an improved convolutional neural network (DC-CNN), comprehensive extraction of global and local features is achieved, and an end-to-end mapping of ground-based cloud images and irradiance is constructed to achieve high-precision prediction of direct solar irradiance.

[0007] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is:

[0008] A hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images includes the following steps:

[0009] Step 1: Obtain historical ground-based cloud images and historical irradiance data and pre-process them, and calculate the corresponding historical solar zenith angle data for model prediction;

[0010] Step 2: Using the knowledge-based cloud detection feature extraction module, based on the red-to-blue ratio (RBR) threshold segmentation method, the influence of aerosol optical depth is considered, and the clear sky library method is used to further extract thin clouds and thick clouds, and obtain a segmented cloud map that distinguishes thin clouds, thick clouds, and clear sky;

[0011] Step 3: Based on the segmented cloud image, a thick and thin cloud matrix U is obtained to distinguish thick clouds, thin clouds, and clear sky pixels, and cloud amount features are extracted according to the proportion of cloud pixels;

[0012] Step 4: further refine the description of cloud amount. Different cloud pixels are assigned different weights according to the segmented cloud image. Combined with the brightness matrix, the cloud thickness coefficient feature is obtained. The knowledge-based cloud detection module is responsible for extracting global features, including cloud amount and cloud thickness coefficient features.

[0013] Step 5, using an improved convolutional neural network (DC-CNN) feature extraction module to better extract local features of the cloud layer and increase the comprehensiveness of cloud image features; the improved convolutional neural network combines dilated convolution with coordinate attention mechanism;

[0014] Step 6: After the two feature extraction modules complete the extraction of global features and local features, the feature fusion and irradiance prediction module is used to comprehensively process the features using a multi-layer fully connected network and a nonlinear activation function (ReLU function), thereby learning the complex relationship between the features. Finally, the predicted value of direct irradiance is obtained through the regression output layer.

[0015] As a preference, the final values ​​of the thin cloud threshold and the thick cloud threshold are determined through experience and repeated tests. The thick and thin cloud matrix U that distinguishes thick clouds, thin clouds and clear sky pixels is obtained according to the final values, and the cloud amount f is calculated according to the following formula:

[0016]

[0017]

[0018] Where m and n represent the pixel size of the image, i and j represent the pixel coordinates in the image, i represents the row index in the image, j represents the column index in the image, Cloud(i,j) and Clear(i,j) are divided into 0 or 1 according to the classification of the pixel, and pixel(i,j) represents the type of pixel at position (i,j) in the image.

[0019] Preferably, in step 4, considering that the clouds in the ground-based cloud map are generally white, and thicker clouds are relatively dark, the cloud map is analyzed by HSV color model, and the V value (brightness value V) representing the image brightness is extracted as a direct indicator for measuring the sky brightness:

[0020] V=max(R,G,B) (4)

[0021] According to the thick and thin cloud matrix U, different weight coefficients are assigned to the corresponding positions of thin cloud, thick cloud and clear sky pixel points, and a matrix with the same dimension as the cloud image pixel is established, with a size of m×n, where m and n are the pixel sizes of the image respectively. The cloud amount matrix W with a refined description is obtained. Combined with the clear sky library method, the V value that best matches the clear sky cloud image at the current moment is extracted, and the brightness matrix V is constructed, with a matrix size of m×n.

[0022] Each cloud image (if the size is 227*227) corresponds to a pixel matrix (the size is also 227*227). The matrix U refers to a matrix that is converted into a matrix for distinguishing thick cloud, thin cloud and clear sky pixel points based on the pixel matrix corresponding to the original cloud image after the aforementioned steps (the original pixel range of each pixel point is 0-255. After the conversion, the original pixel point values ​​are divided into a matrix of three fixed values ​​according to the three different sky types: clear sky, thin cloud and thick cloud).

[0023] The cloud cover matrix W is a matrix obtained by directly assigning different weight coefficients to three different types of pixels on the matrix U, and its size is the same as U.

[0024] By calculating the V value of each pixel point according to the formula of the brightness value V at each pixel position, the V value of this pixel position is taken as the V value of this pixel position. All pixel positions adopt the above processing method to form a brightness matrix V together.

[0025] Furthermore, in order to reflect the impact of clouds on irradiance under different atmospheric aerosol concentrations, the following cloud thickness coefficient calculation formula is used: The cloud thickness of any pixel in the ground-based cloud map can be regarded as the multiplication of the brightness (considering the influence of aerosols) and cloud amount in an ideal cloudless state. CTAs shown below:

[0026]

[0027] Among them, Tr represents the trace of the matrix, T represents the transpose sign, V represents the brightness matrix in an ideal cloud-free state, which is obtained by using the best matching clear sky cloud map at the current moment, W is the cloud amount matrix described in detail, and υ ij is the V value of the pixel with coordinate (i, j) in the matrix V.

[0028] Preferably, in step 5, the improved convolutional neural network (deep neural network) combines dilated convolution (DC) and coordinate attention mechanism (CA) to enhance the ability to capture multi-scale features. Dilated convolution expands the receptive field by introducing intervals in the convolution kernel, so that the network can capture cloud feature information in a larger range without significantly increasing the computational complexity. At the same time, the coordinate attention mechanism effectively enhances the model's attention to features at different positions in the feature map by introducing spatial coordinate information, because it not only pays attention to the feature relationship between channels, but also combines the spatial position of the features, so it performs well in processing tasks involving complex spatial structures, and enhances the network's perception of long-distance dependent features of the cloud layer.

[0029] As a preferred method, the proposed DC-CNN feature extraction module consists of 11 layers, including 5 dilated convolution layers, 3 average pooling layers, 1 coordinate attention mechanism layer, tiling layer and fully connected layer. DC-CNN uses dilated convolution to extract more receptive fields from the input tensor, and adopts a hierarchical feature extraction design. In the early stage, a larger convolution kernel is used to extract more spatial information and primary features, such as the edge of the cloud and the large-area cloud structure. The subsequent convolution layers continue to deepen the extraction of more complex and advanced features, and add a coordinate attention mechanism module deep in the network. Combined with the spatial coordinate information and the feature relationship between channels, it improves the ability to capture information on large-span clouds in the image. After the tiling layer and the fully connected layer, the extracted features are obtained.

[0030] Preferably, in step 6, three fully connected layers are used to learn the nonlinear combination of features. The first two fully connected layers use the relu activation function, have 32 and 64 neurons respectively, and the last output layer has 1 neuron to obtain the solar irradiance prediction result.

[0031] Compared with the prior art, the beneficial effects and advantages of the present invention are:

[0032] (1) By combining knowledge-based cloud detection with an improved convolutional neural network feature extraction module, the present invention can simultaneously extract global features of clouds (such as cloud amount and cloud thickness coefficient) and local features (such as cloud edge and morphology), greatly improving the ability to describe complex cloud features. This joint extraction method of global and local features can effectively address the problem of direct solar irradiance prediction under variable weather conditions.

[0033] (2) Traditional DNI prediction methods are unstable when weather conditions change dramatically. However, the proposed method enhances the perception of large-scale cloud features through dilated convolution and coordinate attention mechanism, so that the model can still maintain high prediction accuracy when dealing with rapid changes in clouds. Experiments show that the proposed model shows good robustness and adaptability under complex meteorological conditions, significantly outperforming the existing benchmark model.

[0034] (3) Through comparative experiments, it can be seen that the present invention has significant improvements in nRMSE and nMAE, which are reduced by at least 15.40% and 27.98% respectively compared with the traditional benchmark model. 2 The value has increased by at least 3.5%, indicating that the accuracy of the model in predicting direct solar irradiance has been significantly improved, which can better support the optimal scheduling of photovoltaic power generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a cloud segmentation image generated using the knowledge-based cloud detection module;

[0036] Figure 2 It is a schematic diagram of the overall structure of the improved convolutional neural network DC-CNN;

[0037] Figure 3 It is a bar chart of the overall performance evaluation indicators of four different prediction models;

[0038] Figure 4 It is the direct irradiance prediction map for high-variability days;

[0039] Figure 5 It is a low-variability daily direct irradiance prediction map. DETAILED DESCRIPTION

[0040] The technical scheme of the present invention is further specifically described below through examples, which are provided for the purpose of illustrating the present invention and are not intended to limit the present invention. Based on the examples in this application, all other examples obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0041] A hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images includes the following steps:

[0042] Step 1: Obtain historical ground-based cloud images and historical irradiance data and pre-process them, and calculate the corresponding historical solar zenith angle data for model prediction;

[0043] Step 2: Using the knowledge-based cloud detection feature extraction module, based on the red-to-blue ratio (RBR) threshold segmentation method, the influence of aerosol optical depth is considered, and the clear sky library method is used to further extract thin clouds and thick clouds, and obtain a segmented cloud map that distinguishes thin clouds, thick clouds, and clear sky;

[0044] Step 3: Based on the segmented cloud image, a thick and thin cloud matrix U is obtained to distinguish thick clouds, thin clouds, and clear sky pixels, and cloud amount features are extracted according to the proportion of cloud pixels;

[0045] Step 4: further refine the description of cloud amount. Different cloud pixels are assigned different weights according to the segmented cloud image. Combined with the brightness matrix, the cloud thickness coefficient feature is obtained. The knowledge-based cloud detection module is responsible for extracting global features, including cloud amount and cloud thickness coefficient features.

[0046] Step 5: Use the improved convolutional neural network (DC-CNN) feature extraction module to better extract the local features of the cloud layer and increase the comprehensiveness of the cloud image features;

[0047] Step 6: After the two feature extraction modules complete the extraction of global features and local features, the feature fusion and irradiance prediction module is used to comprehensively process the features using a multi-layer fully connected network and a nonlinear activation function (ReLU function), thereby learning the complex relationship between the features. Finally, the predicted value of direct irradiance is obtained through the regression output layer.

[0048] The final values ​​of thin cloud threshold and thick cloud threshold are determined through experience and repeated experiments. The thick and thin cloud matrix U that distinguishes thick clouds, thin clouds and clear sky pixels is obtained according to the final values. The cloud amount f is calculated according to the following formula:

[0049]

[0050] Where m and n represent the pixel size of the image, i and j represent the pixel coordinates in the image, i represents the row index in the image, j represents the column index in the image, Cloud(i,j) and Clear(i,j) are divided into 0 or 1 according to the classification of the pixel, and pixel(i,j) represents the type of pixel at position (i,j) in the image.

[0051] In step 4, considering that the clouds in the ground-based cloud image are generally white, and the thicker clouds are relatively dark, the HSV color model analysis is performed on the cloud image, and the V value representing the image brightness is extracted as a direct indicator to measure the sky brightness:

[0052] V=max(R,G,B) (4)

[0053] According to the thick and thin cloud matrix U, different weight coefficients are assigned to the corresponding positions of thin cloud, thick cloud and clear sky pixel points, and a matrix with the same dimension as the cloud image pixel is established, with a size of m×n, where m and n are the pixel sizes of the image respectively. The cloud amount matrix W with a refined description is obtained. Combined with the clear sky library method, the pixel point V value that best matches the clear sky cloud image at the current moment is extracted, and the brightness matrix V is constructed, with a matrix size of m×n.

[0054] In order to reflect the impact of clouds on irradiance under different atmospheric aerosol concentrations, the following cloud thickness coefficient calculation formula is used: The cloud thickness of any pixel in the ground-based cloud map can be regarded as the multiplication of brightness (considering the influence of aerosols) and cloud amount and other elements under ideal cloudless conditions. CT As shown below:

[0055]

[0056] Among them, Tr represents the trace of the matrix, T represents the transpose sign, V represents the brightness matrix in an ideal cloud-free state, which is obtained by using the best matching clear sky cloud map at the current moment, W is the cloud amount matrix described in detail, and υ ij is the V value of the pixel with coordinate (i, j) in the matrix V.

[0057] In step 5, the improved convolutional neural network (DC-CNN) combines dilated convolution (DC) and coordinate attention (CA) to enhance the ability to capture multi-scale features. Dilated convolution expands the receptive field by introducing intervals in the convolution kernel, allowing the network to capture cloud feature information in a larger range without significantly increasing the computational complexity. At the same time, the coordinate attention mechanism effectively enhances the model's attention to features at different positions in the feature map by introducing spatial coordinate information. Because it not only pays attention to the feature relationship between channels, but also combines the spatial position of the features, it performs well in processing tasks involving complex spatial structures and enhances the network's perception of long-distance dependent features of the cloud layer.

[0058] The proposed DC-CNN feature extraction module consists of 11 layers, including 5 dilated convolution layers, 3 average pooling layers, 1 coordinate attention mechanism layer, tiling layer and fully connected layer. DC-CNN uses dilated convolution to extract more receptive fields from the input tensor and adopts a hierarchical feature extraction design. In the early stage, a larger convolution kernel is used to extract more spatial information and primary features, such as the edge of the cloud and the large-area cloud structure. The subsequent convolution layers continue to deepen the extraction of more complex and advanced features, and add a coordinate attention mechanism module deep in the network. Combined with the spatial coordinate information and the feature relationship between channels, it improves the ability to capture the information of large-span clouds in the image. After the tiling layer and the fully connected layer, the extracted features are obtained.

[0059] In step 6, three fully connected layers are used to learn the nonlinear combination of features. The first two fully connected layers use the relu activation function, with 32 and 64 neurons respectively, and the final output layer has 1 neuron to obtain the solar irradiance prediction result.

[0060] The ground-based cloud map and solar measurement data of the National Renewable Energy Laboratory (NREL) of the United States are used to verify the effectiveness of the hybrid prediction method of direct irradiance based on multi-feature extraction of ground-based cloud maps in the present invention. For the processing of ground-based cloud maps, the region of interest (ROI) of the RGB image is mainly extracted to remove unnecessary pixels and retain only the pixels of the sky image. The mask is a binary matrix corresponding to the position of the original ground-based cloud map. The pixel values ​​in the mask are only 0 or 1, 1 corresponds to the position of the sky image in the original ground-based cloud map where the edge obstacles are removed, i.e., ROI, and 0 corresponds to the remaining positions of non-interest in the original ground-based cloud map. After multiplying the original ground-based cloud map with the mask, the processed image is obtained, which is scaled for input to the model with a resolution of 227×227 pixels. Data from 61 full days from April 1 to May 31, 2021 (the data set contains multiple weather types such as sunny, cloudy, and overcast days) are selected. After aligning the image with the numerical data, two typical days are selected as the data source for subsequent case studies, and the remaining 4760 data points are available. In order to ensure effective training and reliable verification, the data is randomly divided into training set, verification set and test set. The training data set consists of 2856 data items. The verification set consists of 476 data items. The test set consists of 1428 data items.

[0061] Figure 1The cloud segmentation images generated by the knowledge-based cloud detection method under different weather types are shown. The original ground-based cloud images (a, c, e, g) and the segmentation images obtained by the knowledge-based cloud detection method (b, d, f, h) are shown: (1) sunny (a, b); (2) few clouds (c, d); (3) cloudy (e, f); (4) cloudy (g, h). In the cloud segmentation images, dark gray indicates thick clouds, light gray indicates thin clouds, and dark blue indicates clear sky. Based on the segmented cloud images, the two global features of cloud amount and cloud thickness coefficient are calculated according to formulas (1), (2), (3) and (4), (5), respectively, which provide important support for the establishment of subsequent models.

[0062] Aiming at local feature extraction of sky images, an improved convolutional neural network DC-CNN feature extraction module was proposed, which combined dilated convolution to expand the receptive field and coordinate attention mechanism to enhance long-range dependencies. Figure 2 This is a schematic diagram of the overall structure of DC-CNN. The colors represent different layer types. The upper part shows the layers of DC-CNN, and the lower part shows the detailed hyperparameters of each layer. The ground cloud map is used as the input of DC-CNN. Through the hierarchical feature extraction design of DC-CNN, the cloud map features are extracted step by step. After the expansion convolution layer, average pooling layer, coordinate attention mechanism layer, and tiling layer, the local features are extracted through the fully connected layer.

[0063] The global and local features extracted by the knowledge-based cloud detection feature extraction module and the improved convolutional neural network (DC-CNN) feature extraction module are used as the input of the feature fusion and irradiance prediction module, and together with the solar altitude angle as input data, the features are processed comprehensively through a multi-layer fully connected network to learn the complex relationship between the features. Finally, the predicted value of direct irradiance is given through the regression output layer.

[0064] In order to verify the contribution of different modules to the overall model performance, the present invention conducted multiple groups of ablation experiments to evaluate the impact of the knowledge-based cloud detection feature extraction module, the DC-CNN feature extraction module and the combination of the two on the final accuracy of the model. The experimental results are shown in Table 1, where (1) Baseline: The baseline model proposed in the present invention contains the main part of the proposed method, but removes the DC and AC components in the DC-CNN feature extraction module and the knowledge-based feature extraction module. (2) Baseline+A: It means that the complete DC-CNN feature extraction module is restored based on the baseline model. (3) Baseline+B: It means that the knowledge-based feature extraction module is embedded in the baseline model. (4) Baseline+A+B: The knowledge-based feature extraction module and the complete DC-CNN feature extraction module, including AC and DC modules, are simultaneously applied in the baseline model.

[0065] Table 1 Ablation experiment results on the test set

[0066] Model <![CDATA[R 2 ]]> nB nMAE nRMSE Baseline 0.8657 -0.1452 0.1987 0.3935 Baseline+A 0.8979 0.0321 0.1731 0.3430 Baseline+B 0.9005 0.0775 0.1645 0.3386 Baseline+A+B 0.9218 0.0098 0.1377 0.3000

[0067] After introducing the knowledge-based cloud detection feature extraction module, the R 2 The DC-CNN feature extraction module improves the R 2 The improvement is 3.72%, and the nRMSE and nMAE are reduced by 12.83% and 12.88% respectively.

[0068] In order to compare with the proposed hybrid model, three classic model methods for solar radiation prediction based on ground-based cloud images are selected as benchmark models. They include (1) SolarNet model, which is based on the VGG-16 architecture and captures the potential relationship between sky images and irradiance. Only a single image is needed to achieve end-to-end irradiance prediction. (2) 3D-CNN model, which models the dynamic changes and spatial distribution of clouds, captures the dynamic characteristics of clouds changing over time, and realizes the prediction of irradiance. (3) CNN-MLP model, which combines convolutional neural network with multi-layer perceptron (MLP), and integrates the cloud features extracted by CNN with other meteorological variables to improve the accuracy of irradiance prediction. The experimental results are shown in Table 2.

[0069] Table 2 Statistical results of overall performance evaluation indicators of four different prediction models

[0070]

[0071]

[0072] The corresponding overall performance index bar chart is drawn as follows Figure 3 As shown in the figure, it can be found that the overall performance of the hybrid model proposed in the present invention is the most superior. Among all the methods, the nRMSE and nMAE values ​​of the method proposed in the present invention are the smallest, which are at least reduced by 15.40%, 27.98% and R 2 The value is the largest, which is at least improved by 3.5%. The corresponding nRMSE is 0.300, nMAE is 0.1377, and R 2 It is 0.9218.

[0073] The model is case validated on a typical day with high variability.

[0074] A high variability day is one in which a high frequency of irradiance changes suddenly occurs after a period of clear skies. Figure 4A high variability daily direct irradiance forecast is shown. This case is challenging due to rapid and sometimes difficult to estimate changes in DNI. The 3D CNN model performs worst in this case, with nRMSE 3D CNN =0.3397, nRMSE given by CNN-MLP model and SolarNet model CNN-MLP =0.2498, nRMSE SolarNet =0.2697; the model proposed in the present invention performs best when used, nRMSE HM =0.1647, nMAE HM =0.1362, R 2 HM =0.8916. In this complex cloud weather situation, the model proposed in the present invention accurately estimates the sudden drop in DNI caused by clouds.

[0075] The model is case validated on a typical day with low variability.

[0076] A low variability day is defined as a day with mostly clear skies, with some clouds appearing during certain times of the day. Figure 5 A low variability daily direct irradiance forecast is shown. The CNN-MLP model performs the worst in this case, with nRMSE CNN-MLP =0.1754, nRMSE given by 3D CNN model and SolarNet model 3DCNN =0.1425, nRMSE SolarNet =0.1313, the model proposed in this invention obtains the best result among the four methods, nRMSE HM =0.0962,nMAE HM =0.0709,R 2 HM =0.9390. From the sky images taken at 14:10 and 14:20, it can be seen that the sun is blocked by the approaching thin clouds, and the DNI drops sharply. Among the four methods, the proposed model accurately captures the drop in DNI, proving the effectiveness of the model.

[0077] The present invention proposes a hybrid prediction model for direct irradiance based on multi-feature extraction of ground-based cloud images, which combines knowledge-based cloud detection with an improved deep convolutional neural network, and fully considers and addresses the problem of direct solar radiation prediction under complex meteorological conditions. By jointly extracting global and local features, the model can more comprehensively describe cloud features, thereby achieving high-precision irradiance prediction. The experimental results verify the effectiveness and superiority of the present invention, especially showing high prediction accuracy and stability under complex and changeable weather conditions.

[0078] The present invention is described in detail above in conjunction with the embodiments, but the contents described are only preferred embodiments of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of application of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images, characterized in that The following steps are involved: Step 1: Obtain historical ground-based cloud images and historical irradiance data and pre-process them, and calculate the corresponding historical solar zenith angle data for model prediction; Step 2: Using the knowledge-based cloud detection feature extraction module, based on the red-to-blue ratio (RBR) threshold segmentation method, the influence of aerosol optical depth is considered, and the clear sky library method is used to further extract thin clouds and thick clouds, and obtain a segmented cloud map that distinguishes thin clouds, thick clouds, and clear sky; Step 3: Based on the segmented cloud image, a thick and thin cloud matrix U is obtained to distinguish thick clouds, thin clouds, and clear sky pixels, and cloud amount features are extracted according to the proportion of cloud pixels; Step 4: further refine the description of cloud amount. Different cloud pixels are assigned different weights according to the segmented cloud image. Combined with the brightness matrix, the cloud thickness coefficient feature is obtained. The knowledge-based cloud detection module is responsible for extracting global features, including cloud amount and cloud thickness coefficient features. Step 5, using an improved convolutional neural network feature extraction module to extract local features of the cloud layer; the improved convolutional neural network combines dilated convolution with a coordinate attention mechanism; Step 6: After completing the extraction of global features and local features, the feature fusion and irradiance prediction module is used, a multi-layer fully connected network is adopted, and the features are comprehensively processed using nonlinear activation functions to learn the complex relationship between the features. Finally, the predicted value of direct irradiance is obtained through the regression output layer.

2. The hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images according to claim 1 is characterized in that: In step 3, the final values ​​of thin cloud threshold and thick cloud threshold are determined through experience and repeated experiments. The thick and thin cloud matrix U that distinguishes thick clouds, thin clouds and clear sky pixels is obtained according to the final values, and the cloud amount f is calculated according to the following formula: Where m and n represent the pixel size of the image, i and j represent the pixel coordinates in the image, i represents the row index in the image, j represents the column index in the image, Cloud(i,j) and Clear(i,j) are divided into 0 or 1 according to the classification of the pixel, and pixel(i,j) represents the type of pixel at position (i,j) in the image.

3. The hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images according to claim 1 is characterized in that: In step 4, the cloud image is analyzed using the HSV color model, and the V value representing the image brightness is extracted as a direct indicator of the sky brightness: V=max(R,G,B) (4) According to the thick and thin cloud matrix U, different weight coefficients are assigned to the corresponding positions of thin cloud, thick cloud and clear sky pixel points, and a matrix with the same dimension as the cloud image pixel is established, with a size of m×n, where m and n are the pixel sizes of the image respectively. The cloud amount matrix W with a refined description is obtained. Combined with the clear sky library method, the V value that best matches the clear sky cloud image at the current moment is extracted, and the brightness matrix V is constructed, with a matrix size of m×n.

4. The hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images according to claim 1 is characterized in that Cloud thickness coefficient characteristic χ CT As shown below: Among them, Tr represents the trace of the matrix, T represents the transpose symbol, V represents the brightness matrix, W is the cloud amount matrix described in detail, and υ ij is the V value of the pixel with coordinate (i, j) in the matrix V.

5. The hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images according to claim 1 is characterized in that: In step 5, the improved convolutional neural network feature extraction module consists of 11 layers, including 5 dilated convolutional layers, 3 average pooling layers, 1 coordinate attention mechanism layer, tiling layer and fully connected layer.

6. The hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images according to claim 5 is characterized in that: The improved convolutional neural network uses dilated convolution to extract more receptive fields from the input tensor and adopts a hierarchical feature extraction design. In the early stage, larger convolution kernels are used to extract more spatial information and primary features. The subsequent convolution layers continue to deepen the extraction of more complex and advanced features, and add a coordinate attention mechanism module deep in the network. It combines the spatial coordinate information and the feature relationship between channels to improve the ability to capture information on large-span clouds in the image. After the tiling layer and the fully connected layer, the extracted features are obtained.

7. The hybrid prediction method for direct irradiance based on multi-feature extraction of ground-based cloud images according to claim 1 is characterized in that: In step 6, three fully connected layers are used to learn nonlinear combinations of features.

8. The direct irradiance hybrid prediction method based on multi-feature extraction of ground-based cloud images according to claim 7 is characterized by: The first two fully connected layers use nonlinear activation functions and have 32 and 64 neurons respectively. The final output layer has 1 neuron to obtain the solar irradiance prediction results.

Images