Photovoltaic battery control method and device based on environmental perception

Abstract

Embodiments of the present disclosure disclose an environment-aware photovoltaic battery control method and device. A specific embodiment of the method comprises: acquiring a sequence of sky images; performing cumulus cloud identification on each sky image in the sequence of sky images according to target region corresponding wind state information; determining occlusion description information according to the sequence of cumulus cloud description information and the sun position information; in response to the occlusion description information satisfying a trigger condition, performing discharge control on the photovoltaic batteries in the photovoltaic battery array according to the occlusion description information and load description information; and in response to the occlusion description information not satisfying the trigger condition, performing charging control on the photovoltaic batteries in the photovoltaic battery array according to the load description information. The embodiment reduces the equipment cost, and effectively reduces the attenuation speed of the battery through effective battery charging and discharging control.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the fields of computer technology, image processing, battery charging and discharging control, and photovoltaic power generation, specifically to a photovoltaic battery control method and apparatus based on environmental perception. Background Technology

[0002] Photovoltaic power generation is a clean power generation technology that converts solar energy into electrical energy, currently relying primarily on photovoltaic panels based on crystalline silicon materials. The impact of cumulus clouds on photovoltaic power generation is complex and significant. Specifically, when cumulus clouds drift between the sun and the photovoltaic panel, they not only directly affect power generation but also cause short-term fluctuations in power output, thus impacting power generation stability. Conventional methods primarily employ power equipment such as inverters (e.g., microinverters) for instantaneous control.

[0003] However, when using the above method, the following technical problems often arise: A large number of inverters need to be set up for independent control, which leads to complex control logic and also greatly increases the cost of equipment.

[0004] Therefore, based on this, the industry has proposed some battery-based solutions to address the instantaneous power changes caused by cumulus clouds. However, these technical solutions have failed to effectively and accurately control battery charging and discharging, which accelerates battery aging in scenarios with frequent cumulus clouds, thus greatly increasing the rate of battery degradation. Summary of the Invention

[0005] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0006] Some embodiments of this disclosure propose a photovoltaic battery control method and apparatus based on environmental perception to solve the technical problems mentioned in the background section above.

[0007] In a first aspect, some embodiments of this disclosure provide a photovoltaic battery control method based on environmental perception. The method includes: acquiring a sequence of sky images, wherein the sky images correspond to the sky above a target area, the target area being an area where a photovoltaic panel array is installed; identifying cumulus clouds in each sky image in the sky image sequence based on wind state information corresponding to the target area, obtaining a sequence of cumulus cloud description information, wherein the cumulus cloud description information includes: cumulus cloud position, cumulus cloud movement speed, and cumulus cloud confidence; determining shading description information based on the cumulus cloud description information sequence and solar position information, wherein the shading description information includes: predicted shading start time, predicted shading duration, and shading confidence; responding to the shading description information satisfying a trigger condition, discharging the photovoltaic batteries in the photovoltaic battery array based on the shading description information and load description information, wherein the photovoltaic batteries store electricity through the photovoltaic panel array; and responding to the shading description information not satisfying the trigger condition, storing electricity in the photovoltaic batteries in the photovoltaic battery array based on the load description information.

[0008] Secondly, some embodiments of this disclosure provide a photovoltaic battery control device based on environmental perception. The device includes: an acquisition unit configured to acquire a sequence of sky images, wherein the sky images correspond to the sky above a target area, and the target area is an area where a photovoltaic panel array is installed; a cumulus cloud identification unit configured to perform cumulus cloud identification on each sky image in the sky image sequence based on wind state information corresponding to the target area, to obtain a cumulus cloud description information sequence, wherein the cumulus cloud description information includes: cumulus cloud location, cumulus cloud movement speed, and cumulus cloud confidence level; and a determination unit configured to determine the cumulus cloud description information sequence and... The system includes: solar position information; shading description information, including predicted shading start time, predicted shading duration, and shading confidence level; a discharge control unit configured to, in response to the shading description information meeting the triggering conditions, control the discharge of photovoltaic cells in the photovoltaic cell array based on the shading description information and load description information, wherein the photovoltaic cells store electricity through the photovoltaic panel array; and a storage control unit configured to, in response to the shading description information not meeting the triggering conditions, control the storage of photovoltaic cells in the photovoltaic cell array based on the load description information.

[0009] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0010] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0011] The above embodiments of this disclosure have the following beneficial effects: The photovoltaic battery control method based on environmental perception, as described in some embodiments of this disclosure, reduces equipment costs and, through effective battery charging and discharging control, reduces the rate of battery degradation. Specifically, firstly, a sequence of sky images is acquired, where the sky images correspond to the sky above a target area, which is the area where the photovoltaic panel array is located. In practice, cumulus clouds are one of the main causes of short-term power generation changes; therefore, sky images are acquired for subsequent cumulus cloud identification. Secondly, based on the wind state information corresponding to the target area, cumulus cloud identification is performed on each sky image in the above sky image sequence to obtain a cumulus cloud description information sequence, where the cumulus cloud description information includes: cumulus cloud location, cumulus cloud movement speed, and cumulus cloud confidence level. Through cumulus cloud identification, the distribution and changes of cumulus clouds above the photovoltaic panel array are obtained. Next, based on the above cumulus cloud description information sequence and solar position information, shading description information is determined, where the shading description information includes: predicted shading start time, predicted shading duration, and shading confidence level. In practice, when cumulus clouds are located between the sun and photovoltaic (PV) panels, they can affect the instantaneous power generation of the PV panels. Therefore, this disclosure combines cumulus cloud description information sequences and solar position information to determine the start time, duration, and corresponding confidence level of shading through shading prediction. Furthermore, in response to the aforementioned shading description information meeting the triggering conditions, the PV batteries in the PV battery array are discharged based on the aforementioned shading description information and load description information, wherein the PV batteries store electricity through the PV panel array. By combining the shading description information and load description information, the discharge control of the PV batteries is achieved, thereby reducing the impact of cumulus clouds on the instantaneous power generation of PV power, thus ensuring the stable operation of the load equipment. Finally, in response to the aforementioned shading description information not meeting the triggering conditions, the PV batteries in the aforementioned PV battery array are charged based on the aforementioned load description information. This achieves intelligent charging control of PV batteries while ensuring the stable operation of the load equipment. This method eliminates the need for a large number of inverters and other power equipment, thereby reducing equipment costs, and effectively controlling battery charging and discharging reduces the rate of battery degradation. Attached Figure Description

[0012] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0013] Figure 1 This is a flowchart of some embodiments of the photovoltaic battery control method based on environmental perception according to the present disclosure; Figure 2 This is a schematic diagram of the model architecture for cumulus cloud recognition; Figure 3 This is a schematic diagram of the module structure of the backbone feature extraction module; Figure 4 This is a schematic diagram of the module structure of the multi-scale feature extraction module; Figure 5 This is a schematic diagram of the structure of some embodiments of the photovoltaic battery control device based on environmental perception according to the present disclosure; Figure 6 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0014] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0015] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0016] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0017] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0018] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0019] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] refer to Figure 1 The diagram illustrates a flow 100 of some embodiments of the environment-aware photovoltaic battery control method according to the present disclosure. This environment-aware photovoltaic battery control method includes the following steps: Step 101: Obtain the sky image sequence.

[0021] In some embodiments, the executor of the environment-aware photovoltaic battery control method (e.g., a computing device) can acquire a sequence of sky images via a wired or wireless connection.

[0022] The sky image corresponds to the sky above the target area. The target area is the region where the photovoltaic panel array is located. The sky images in the sky image sequence can be continuously acquired by a full-sky imager. The photovoltaic panel array can contain multiple photovoltaic panels. Specifically, assuming the area of ​​the target area is S, and the projection range of the sky above the target area onto the ground corresponding to the full-sky imager is N, then S∈N. The sky image is an RGB (Red, Green, Blue) image.

[0023] In practice, since photovoltaic panels generate electricity through solar energy conversion, the effective power generation period is from sunrise to sunset. Furthermore, the image quality of all-sky imagers is poor at night. Therefore, all-sky imagers can acquire sky images during this period. Additionally, when the weather information for the target area indicates rain, snow, fog, or haze, image acquisition by the all-sky imager can be stopped to optimize its image acquisition logic and improve the quality of the resulting sky images.

[0024] It should be noted that the aforementioned computing devices can be either hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or a single terminal device. When the computing device is software, it can be installed on the hardware devices listed above. It can be implemented as, for example, multiple software programs or software modules used to provide distributed services, or as a single software program or software module. No specific limitations are made here.

[0025] It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (Ultra Wide Band) connections, and other currently known or future wireless connection methods.

[0026] Step 102: Based on the wind state information corresponding to the target area, perform cumulus cloud identification on each sky image in the sky image sequence to obtain a cumulus cloud description information sequence.

[0027] In some embodiments, the aforementioned execution entity can identify cumulus clouds in each sky image in the sky image sequence based on the wind state information corresponding to the target area, thereby obtaining a sequence of cumulus cloud description information.

[0028] The wind state information represents the wind speed and direction corresponding to the target area. Specifically, wind state information can be represented using wind vectors. The direction of the wind vector represents the wind direction. The magnitude of the wind vector represents the wind speed. The cumulus cloud description information represents the cumulus cloud state within the sky image. This information includes: cumulus cloud location, cumulus cloud movement speed, and cumulus cloud confidence score. The cumulus cloud location represents the position of the identified cumulus cloud in the sky above the target area. The cumulus cloud movement speed represents the movement speed of the identified cumulus cloud. The cumulus cloud confidence score represents the confidence level of the identified cumulus cloud.

[0029] In practice, firstly, object detection methods can be used to identify cumulus clouds in sky images to obtain cumulus cloud description information. For example, the YOLOv3 (You Only Look Once Version 3) model can be used for cumulus cloud identification in sky images. Specifically, when using the YOLOv3 model for cumulus cloud identification, the identified coordinates are two-dimensional coordinates in the image coordinate system. Therefore, before using an all-sky imager, corresponding camera calibration is required to obtain the corresponding rotation matrix and translation vector. Based on this, given the two-dimensional coordinates of the cumulus clouds, the rotation matrix and translation vector are combined to convert the obtained two-dimensional coordinates into three-dimensional coordinates in the geodetic coordinate system, which serve as the cumulus cloud location included in the cumulus cloud description information. Simultaneously, the confidence score of the cumulus clouds identified by the YOLOv3 model can be used as the confidence score of the cumulus cloud included in the cumulus cloud description information. Furthermore, since the sky images in the sky image sequence are sampled at a fixed frequency (e.g., 10 images per second),... Therefore, the cumulus cloud movement speed can be determined based on the cumulus cloud location information and the acquisition time difference between two adjacent sky images, where the cumulus cloud movement speed = cumulus cloud location difference / acquisition time difference. In particular, considering that the movement of cumulus clouds at high altitudes is affected by wind speed and direction, the movement direction of the convolution operator in the YOLOv3 model during image convolution processing can be adjusted by incorporating wind state information. Specifically, since the wind vector is a vector in the geodetic coordinate system, it can first be transformed into the image coordinate system using a rotation matrix and a translation vector. Next, the direction angle between the wind vector and the horizontal direction in the image coordinate system, and the direction angle between the wind vector and the vertical direction in the image coordinate system, are determined. Furthermore, when the direction angle between the wind vector and the horizontal direction in the image coordinate system is smaller than the direction angle between the wind vector and the vertical direction in the image coordinate system, the movement direction of the convolution operator is adjusted to a horizontal movement. When the wind vector in the image coordinate system has a horizontal direction angle greater than or equal to the vertical direction angle, the convolution operator is moved vertically. This, combined with wind state information, enables directional convolution feature extraction, resulting in more distinct cumulus cloud boundary features.

[0030] In some optional implementations of certain embodiments, the execution entity performs cumulus cloud identification on each sky image in the sky image sequence based on the wind state information corresponding to the target area, to obtain a cumulus cloud description information sequence, including: Step S1: Perform color space conversion on the above sky image to obtain the converted sky image.

[0031] The converted sky image is an HSV (Hue, Saturation, Value) image.

[0032] As an example, color space conversion can be performed using the following code: `hsv_img = cv2.cvtColor(rgb_img, cv2.COLOR_RGB2HSV)`, where "rgb_img" represents the sky image, and "hsv_img" represents the converted sky image.

[0033] Step S2: Perform image enhancement on the converted sky image to obtain the enhanced sky image.

[0034] The enhanced sky image is an image that has undergone brightness enhancement and saturation suppression.

[0035] In practice, cumulus clouds typically reflect a large amount of light, thus exhibiting higher brightness relative to the background. Simultaneously, cumulus clouds are often white or off-white, exhibiting low saturation. Therefore, to highlight cumulus clouds, this disclosure employs brightness enhancement and saturation suppression. Specifically, brightness enhancement emphasizes the brightness difference between cumulus clouds and the background, while saturation suppression emphasizes the saturation difference between them. Specifically, corresponding enhancement and suppression coefficients can be set for brightness and saturation, respectively. Since the converted sky image is an HSV image, the brightness of the V (Value) channel of the converted sky image can be enhanced using an enhancement coefficient, i.e., V channel × enhancement coefficient, where the value range of the enhancement coefficient is (1, 1.4). Similarly, the saturation of the S (Saturation) channel of the converted sky image can be suppressed using a suppression coefficient, i.e., S channel × suppression coefficient, where the value range of the suppression coefficient is [0.6, 1].

[0036] Step S3: Using a pre-trained cumulus recognition model, perform cumulus recognition on the enhanced sky image to obtain the cumulus description information corresponding to the sky image, including the cumulus location and cumulus confidence level.

[0037] The cumulus cloud recognition model is a machine learning model used to identify cumulus clouds in enhanced sky images. For example, the cumulus cloud recognition model can use the YOLOv3 model.

[0038] In practice, firstly, when using the YOLOv3 model for cumulus cloud identification, the identified coordinates are two-dimensional coordinates in the image coordinate system. Therefore, before using an all-sky imager, corresponding camera calibration is required to obtain the corresponding rotation matrix and translation vector. Based on this, and using the rotation matrix and translation vector, the obtained two-dimensional coordinates are converted into three-dimensional coordinates in the geodetic coordinate system, which serve as the cumulus cloud location included in the cumulus cloud description information. Furthermore, the confidence score of the cumulus clouds identified by the YOLOv3 model can be used as the confidence score of the cumulus clouds included in the cumulus cloud description information.

[0039] Optionally, the cumulus cloud recognition model includes: a backbone feature extraction module, a multi-scale feature extraction module, a morphological feature recognition module, and a detection head group. The morphological feature recognition module includes: a cumulus cloud edge feature extractor, a cumulus cloud texture feature extractor, and a cumulus cloud color feature extractor. The detection head group consists of a first detection head, a second detection head, and a third detection head. The first detection head is used for cumulus cloud boundary point localization. The second detection head is used for cumulus cloud center point localization. The third detection head is used for cumulus cloud confidence generation.

[0040] As an example, see Figure 2 The diagram illustrates the architecture of the cumulus cloud recognition model. The core feature extraction module, multi-scale feature extraction module, morphological feature recognition module, and detection head group are connected in series. The input to the core feature extraction model is the enhanced sky image. The input to the multi-scale feature extraction module is the output of the core feature extraction model. The input to the morphological feature recognition module is the output of the multi-scale feature extraction module. The input to the detection head group is the output of the morphological feature recognition module. Specifically, the morphological feature recognition module includes a parallel cumulus cloud edge feature extractor, a cumulus cloud texture feature extractor, and a cumulus cloud color feature extractor. The cumulus cloud edge feature extractor is used for cumulus cloud edge feature extraction, the cumulus cloud texture feature extractor is used for cumulus cloud texture feature extraction, and the cumulus cloud color feature extractor is used for cumulus cloud color feature extraction. The first detection head, the second detection head, and the third detection head in the detection head group are configured in parallel. Specifically, the cumulus cloud recognition model adopts the aforementioned architecture for the following reasons: First, to meet the cumulus cloud recognition needs under different computing power scenarios, the model employs a decoupled network design, using four decoupled modules (backbone feature extraction module, multi-scale feature extraction module, morphological feature recognition module, and detection head group). This allows the network structure to be adjusted according to different computing power constraints. Second, cumulus clouds are complex three-dimensional structures formed by the aggregation of tiny water droplets or ice crystals, thus exhibiting irregular and blurred boundary features in the boundary dimension. Third, cumulus clouds are often composed of multiple small cloud clusters, resulting in overlapping textures of varying depths in the texture dimension. Furthermore, cumulus clouds exhibit low saturation and high brightness, and the enhanced sky image also undergoes brightness enhancement and saturation suppression to highlight the color difference between the cumulus clouds and the background, thereby presenting a more prominent brightness and saturation expression in the color dimension. Building upon this, the morphological feature extraction module extracts directional features in three specific dimensions—cumulus edge features, cumulus texture features, and cumulus color features—to improve the effectiveness of feature extraction. Finally, the module outputs cumulus description information, including cumulus location, cumulus movement speed, and cumulus confidence level, through three detection heads.

[0041] In some optional implementations of certain embodiments, the execution entity performs cumulus cloud recognition on the enhanced sky image using a pre-trained cumulus cloud recognition model to obtain cumulus cloud description information corresponding to the sky image, including cumulus cloud location and cumulus cloud confidence, including: Step S31: Perform coarse feature extraction on the enhanced sky image using the above-mentioned backbone feature extraction module to obtain the initial sky image features.

[0042] As an example, see Figure 3 The diagram shows the module structure of the backbone feature extraction module, which includes: convolutional layer, normalization layer, ReLU activation function, max pooling layer, residual block R1, residual block R2, residual block R3, residual block R4, residual block R5, residual block R6, residual block R7, and residual block R8.

[0043] The convolutional layers have a 7×7 kernel size and a stride of 2. Batch normalization (BN) is used for the normalization layers. The max-pooling layers have a 3×3 pooling window size and a stride of 2. The output of the max-pooling layers corresponds to a feature dimension of (H / 4, W / 4, 64). Residual block R1 has a 3×3 kernel size and a stride of 1. Residual block R2 has a 3×3 kernel size and a stride of 1. The output of residual block R2 corresponds to a feature dimension of (H / 4, W / 4, 128). Residual block R3 has a 3×3 kernel size and a stride of 2. Residual block R4 has a 3×3 kernel size and a stride of 1. The output of residual block R4 corresponds to a feature dimension of (H / 8, W / 8, 256). Residual block R5 has a 3×3 kernel size and a stride of 2. The convolutional kernel size of residual block R6 is 3×3 with a stride of 1. The feature dimension corresponding to the output of residual block R6 is (H / 16, W / 16, 512). The convolutional kernel size of residual block R7 is 3×3 with a stride of 2. The convolutional kernel size of residual block R8 is 3×3 with a stride of 1. The feature dimension corresponding to the output of residual block R8 is (H / 32, W / 32, 1024). The outputs of residual blocks R2, R4, R6, and R8 together constitute the initial sky image features.

[0044] Step S32: Perform multi-scale feature extraction based on the multi-scale feature extraction module and the initial sky image features to obtain a set of multi-scale sky image features.

[0045] The multi-scale sky image feature set consists of a first multi-scale sky image feature and a second multi-scale sky image feature. The first multi-scale sky image feature is extracted using a horizontal convolution operator, and the second multi-scale sky image feature is extracted using a vertical convolution operator. Both the horizontal and vertical convolution operators are one-dimensional convolution operators.

[0046] As an example, see Figure 4The diagram shows the module structure of the multi-scale feature extraction module, which consists of a feature shaping module, multi-scale feature extraction branches M1, M2, and M3. Taking an initial sky image feature as an example, the corresponding multi-scale sky image feature set includes three multi-scale sky image feature sets. Specifically, firstly, since the initial sky image feature is composed of the outputs of residual blocks R2, R4, R6, and R8, the feature shaping module includes four parallel 1×1 convolutional layers and three upsampling layers to perform parallel feature channel number shaping on four features with feature dimensions (H / 4, W / 4, 128), (H / 8, W / 8, 256), (H / 16, W / 16, 512), and (H / 32, W / 32, 1024), resulting in four shaped features. The four shaped features have the same number of feature channels, and their corresponding feature dimensions are... The features are (H / 4, W / 4, 256), (H / 8, W / 8, 256), (H / 16, W / 16, 256), and (H / 32, W / 32, 256) respectively. The three shaped features (H / 8, W / 8, 256), (H / 16, W / 16, 256), and (H / 32, W / 32, 256) are upsampled in parallel to (H / 4, W / 4, 256) through three upsampling layers. These features are then superimposed with the shaped features (H / 4, W / 4, 256) to form the concatenated features (feature dimension (H / 4, W / 4, 256)). Secondly, the multi-scale feature extraction branches M1, M2, and M3 each include one deconvolutional layer and one bidirectional convolutional layer, respectively. All three branches take concatenated features with a feature dimension of (H / 4, W / 4, 256) as input. The output of the deconvolutional layer in branch M1 has a feature dimension of (H / 2, W / 2, 256). The output of the deconvolutional layer in branch M2 has a feature dimension of (H, W, 256). The output of the deconvolutional layer in branch M3 has a feature dimension of (2H, 2W, 256). All bidirectional convolutional layers employ both horizontal and vertical convolution operators. The scale feature extraction branch M1 includes a bidirectional convolutional layer, which performs bidirectional convolution (horizontal and vertical) on the output of the multi-scale feature extraction branch M1, which includes a deconvolution layer, through the corresponding horizontal convolution operator and vertical convolution operator, respectively, to obtain the first multi-scale sky image feature and the second multi-scale sky image feature of the multi-scale sky image feature group corresponding to the multi-scale feature extraction branch M1.The scale feature extraction branch M2 includes bidirectional convolutional layers that perform bidirectional convolution (horizontal and vertical) on the outputs of the two deconvolutional layers in the multi-scale feature extraction branch M2 using corresponding horizontal and vertical convolution operators, respectively, to obtain the first and second multi-scale sky image features included in the multi-scale sky image feature set corresponding to the multi-scale feature extraction branch M2. The scale feature extraction branch M3 includes bidirectional convolutional layers that perform bidirectional convolution (horizontal and vertical) on the outputs of the three deconvolutional layers in the multi-scale feature extraction branch M3 using corresponding horizontal and vertical convolution operators, respectively, to obtain the first and second multi-scale sky image features included in the multi-scale sky image feature set corresponding to the multi-scale feature extraction branch M3.

[0047] Step S33: Determine the horizontal and vertical weights based on the wind state information above.

[0048] The horizontal weights correspond to the features of the first multi-scale sky image, while the vertical weights correspond to the features of the second multi-scale sky image.

[0049] In practice, since wind state information is represented by wind vectors, the horizontal weight = angle between the wind vector and the horizontal direction / 90°, and the vertical weight = angle between the wind vector and the vertical direction / 90°. Where 0 ≤ angle between the wind vector and the horizontal direction ≤ 90°, and the sum of the angles between the wind vector and the vertical direction and the horizontal direction equals 90°.

[0050] Step S34: For each multi-scale sky image feature group in the above multi-scale sky image feature group set, according to the above horizontal weight and the above vertical weight, the first multi-scale sky image features and the second multi-scale sky image features included in the above multi-scale sky image feature group are superimposed with feature weights to obtain fused image features.

[0051] Wherein, the fused image features = horizontal weight × first multi-scale sky image features + vertical weight × second multi-scale sky image features.

[0052] Step S35: Perform cumulus cloud morphology recognition based on the above morphology feature recognition module and the obtained fused image feature set to obtain cumulus cloud morphology features.

[0053] The aforementioned cumulus cloud morphological features consist of cumulus cloud edge features, cumulus cloud texture features, and cumulus cloud color features.

[0054] The cloud edge feature extractor consists of deformable convolutional layers, 3×3 convolutional layers, and 1×1 convolutional layers. The cumulus cloud texture feature extractor consists of a Gabor filter bank (horizontal and vertical Gabor filters), 3×3 convolutional layers, and 1×1 convolutional layers. The cumulus cloud color feature extractor consists of a global average pooling layer and a fully connected layer.

[0055] In practice, firstly, the aforementioned executing entity performs feature stitching along channel (256) on the fused image feature set, using these as inputs to the cloud edge feature extractor, cumulus texture feature extractor, and cumulus color feature extractor, respectively, to obtain cumulus edge features, cumulus texture features, and cumulus color features. Next, the cumulus edge features, cumulus texture features, and cumulus color features are stitched together to obtain cumulus morphology features.

[0056] Step S36: Based on the above detection head group and the above cumulus morphology features, generate cumulus description information corresponding to the above sky image, including cumulus location and cumulus confidence.

[0057] The system comprises three detector heads: a first detector head for cumulus boundary point localization, a second detector head for cumulus center point localization, and a third detector head for cumulus confidence generation. The first detector head consists of a 3×3 convolutional layer, a 1×1 convolutional layer, a global average pooling layer, a fully connected layer, and a Tanh activation function. The second detector head has the same structure as the first. The third detector head consists of a 3×3 convolutional layer, a 1×1 convolutional layer, a global average pooling layer, a fully connected layer, and a sigmoid activation function. All three detector heads take cumulus morphology features as input. Specifically, since the outputs of the first and second detector heads are two-dimensional coordinates in the image coordinate system, it is necessary to combine the rotation matrix and translation vector of the all-sky imager to convert the multiple two-dimensional coordinates (two-dimensional coordinates of cumulus boundary points and two-dimensional coordinates of cumulus center points) output by the first and second detector heads into three-dimensional coordinates in the geodetic coordinate system. These coordinates serve as the cumulus location information included in the cumulus description information corresponding to the sky image. Furthermore, the confidence level output by the third detection head can be directly used as the cumulus confidence level included in the cumulus description information corresponding to the sky image.

[0058] Step S4: In response to the fact that the above sky image is the first sky image in the above sky image sequence, the preset speed is determined to be the cumulus movement speed included in the cumulus description information corresponding to the above sky image.

[0059] In practice, the cumulus movement speed is calculated as: Cumulus position difference / Acquisition time difference. Here, the cumulus position difference represents the positional difference between the cumulus locations included in the two cumulus descriptions corresponding to two sky images. The acquisition time difference represents the time difference between the two sky images during the image acquisition process. Since the first sky image does not contain any preceding sky images, a preset speed (e.g., 0) is used as the cumulus movement speed included in the cumulus description information corresponding to the first sky image.

[0060] Step S5: In response to the above sky image not being the first sky image in the above sky image sequence, generate the cumulus movement speed included in the cumulus description information corresponding to the above sky image based on the previous cumulus position and the cumulus position included in the cumulus description information corresponding to the above sky image.

[0061] The foreground cumulus location represents the location of the cumulus clouds included in the cumulus cloud description information corresponding to the sky image preceding the aforementioned sky image.

[0062] In practice, firstly, the difference between the cumulus position and the preceding cumulus position included in the cumulus description information corresponding to the sky image can be used as the cumulus position difference. Then, the time difference between the aforementioned sky image and the sky image corresponding to the preceding cumulus position during the image acquisition process is determined as the acquisition time difference. Finally, the cumulus movement speed included in the cumulus description information corresponding to the aforementioned sky image is determined using the formula: Cumulus movement speed = Cumulus position difference / Acquisition time difference.

[0063] Step 103: Determine the occlusion description information based on the cumulus cloud description information sequence and the sun position information.

[0064] In some embodiments, the aforementioned executing entity can determine the occlusion description information based on the cumulus cloud description information sequence and the sun's position information.

[0065] The solar position information represents the current position of the sun. For example, solar position information may include: solar altitude, solar azimuth, etc. The shading description information includes: predicted shading start time, predicted shading duration, and shading confidence level. The predicted shading start time represents the predicted start time of cumulus cloud shading of the photovoltaic panel array. The predicted shading duration represents the predicted duration of cumulus cloud shading of the photovoltaic panel array. The shading confidence level represents the confidence level that cumulus cloud shading of the photovoltaic panel array will occur.

[0066] In practice, cumulus clouds can cause shading when they are located between the sun and a photovoltaic array. The cumulus cloud description information sequence describes the changes in the cloud's position and speed. Therefore, a time series prediction method can be used to predict the two predicted times when the cumulus cloud moves between the sun and the photovoltaic array and when it moves outside the sun and the photovoltaic array. This gives us the predicted shading start time (the first predicted time) and the predicted shading duration (the time difference between the two predicted times). The average confidence level of the two predicted times is then used as the shading confidence level.

[0067] In some optional implementations of certain embodiments, the execution entity determines the occlusion description information based on the cumulus description information sequence and the sun position information, including: Step S1: Generate the fence space based on the above solar position information and the area position information corresponding to the above target area.

[0068] The aforementioned fence space is a four-sided pyramidal space.

[0069] In practice, the solar position, represented by the solar position information, can be used as the apex of the cone, and the plane where the target area is located can be used as the base of the cone to form a fence space.

[0070] Step S2: Extract temporal location features from the cumulus locations included in the above cumulus description information sequence to generate cumulus location features.

[0071] Step S3: Extract temporal motion features from the cumulus movement speed included in the cumulus description information sequence above to generate cumulus movement speed features.

[0072] In practice, since cumulus location and cumulus movement speed exhibit distinct temporal characteristics, two temporal feature extraction models (e.g., RNN (Recurrent Neural Network) or LSTM) can be used to extract features from the cumulus description information sequence, specifically the cumulus location and cumulus movement speed features. In particular, the two temporal feature extraction models have identical structures to ensure consistency in the feature representation of cumulus location and cumulus movement speed.

[0073] Step S4: Based on the above-mentioned cumulus location characteristics and cumulus movement speed characteristics, determine the predicted location and prediction confidence level.

[0074] In practice, the two temporal feature extraction models share a single prediction head. This involves aligning the cumulus location features and the aforementioned cumulus movement speed features over time to obtain aligned features. (Since the two LSTM models have identical structures, the output cumulus location and movement speed features have consistent feature representations in the time dimension, thus allowing direct alignment of corresponding point feature values.) The prediction head consists of four fully connected layers and a ReLU activation function, recursively predicting the cumulus location at time T+k to obtain the predicted location and corresponding prediction confidence.

[0075] Step S5: In response to the predicted position being within the fence space, perform the following processing steps: Step S51: Determine the average cumulus movement speed based on the cumulus movement speed included in the cumulus description information in the above cumulus description information sequence.

[0076] In practice, since the spatial range of the fence space is fixed, it is possible to determine whether the predicted position is located within the fence space by judging the positional relationship between the predicted position and the four triangular cones of the fence space.

[0077] In practice, the average speed of cumulus movement included in the last three cumulus location description information sequences can be used as the average cumulus speed.

[0078] Step S52: Based on the above average cumulus cloud movement speed, the above predicted position, the above fence space and the target cumulus cloud position, determine the predicted occlusion start time and predicted occlusion duration included in the above occlusion description information.

[0079] The target cumulus location mentioned above refers to the cumulus location included in the cumulus description information located at the end of the sequence of cumulus description information.

[0080] In practice, firstly, a direction vector can be constructed based on the target cumulus cloud location and the predicted location. This direction vector points from the target cumulus cloud location to the predicted location. Then, the intersection points of this direction vector and the triangular pyramids of the fence space are determined (usually two: the location where the cumulus cloud enters the fence space and the location where it exits). Next, the ratio of the distance between the intersection point of the triangular pyramids corresponding to the location entering the fence space and the target cumulus cloud location to the average cumulus cloud movement speed is determined as the time increment. Further, the time increment is added to the image acquisition time of the sky image corresponding to the cumulus cloud location's description information to obtain the predicted occlusion start time. Additionally, the intersection point distance is determined based on the obtained triangular pyramid intersection points. Finally, the ratio of the intersection point distance to the average cumulus cloud movement speed is determined as the predicted occlusion duration.

[0081] Step S53: Determine the prediction confidence as the occlusion confidence included in the occlusion description information.

[0082] Step 104: In response to the shading description information meeting the trigger condition, discharge control is performed on the photovoltaic batteries in the photovoltaic battery array based on the shading description information and the load description information.

[0083] In some embodiments, the aforementioned execution entity may, in response to the shading description information meeting the triggering condition, perform discharge control on the photovoltaic batteries in the photovoltaic battery array based on the shading description information and the load description information.

[0084] The load description information represents the load devices powered by photovoltaic panel arrays and / or photovoltaic battery arrays. The load description information may include: load type and load power. The photovoltaic battery array may be an energy storage device composed of multiple batteries. The triggering condition may be: the predicted shading start time and predicted shading duration are not null values, and the shading confidence level is greater than or equal to a preset confidence level.

[0085] In practice, assume the power output of the photovoltaic (PV) panel array is P1. The load power of the corresponding load device is P2. Where P1 ≥ P2. It is known that cumulus cloud cover causes significant changes in the power output of the PV panel array. Therefore, the PV batteries in the PV battery array can be controlled to discharge for a preset time before the predicted start time of shading. This means that the PV batteries and PV panels simultaneously supply power to the load device to stabilize its load power. Specifically, the discharge power of the PV batteries can be P3, where P3 + α × P1 ≥ P2. Here, α is the proportion of PV panel power attenuation caused by cumulus cloud cover, as measured based on historical data. The PV batteries are then controlled to discharge at a power output of P3 for a specified duration based on the predicted shading duration.

[0086] In some optional implementations of some embodiments, the execution entity, in response to the shading description information satisfying the triggering condition, performs discharge control on the photovoltaic cells in the photovoltaic battery array based on the shading description information and the load description information, including: Step S1: Determine the battery status of the photovoltaic batteries in the photovoltaic battery array to obtain a set of battery status information.

[0087] Among them, battery status information represents the battery status of photovoltaic batteries. Battery status information may include, but is not limited to: battery identification, battery voltage, battery current, battery capacity, and battery temperature.

[0088] In practice, the battery status of each photovoltaic cell in a photovoltaic battery array can be read through a battery management system (BMS) to obtain the corresponding battery status information. The BMS is a system used to monitor the battery status of photovoltaic cells.

[0089] Step S2: Determine the target photovoltaic battery set based on the above load description information, the above battery status information set, the predicted shading duration and the preset power ratio included in the above shading description information.

[0090] The target photovoltaic battery is the photovoltaic battery in the aforementioned photovoltaic battery array that is used to supply power to the applied electrical load based on the load description information.

[0091] In practice, firstly, assume the power output of the photovoltaic (PV) panel array is P1. The load power of the load device corresponding to the load description information is P2. α is the proportion (power percentage) of PV panel power attenuation caused by cumulus clouds, measured based on historical data. The discharge power of the PV battery can be P3. Therefore, it should be ensured that P3 + α × P1 ≥ P2. Taking P3 + α × P1 = P2 as an example, the specific value of P3 can be obtained. Next, assume the PV batteries in the PV battery array are connected in parallel. Therefore, the battery power P4 corresponding to a single PV battery can be determined based on the battery current and battery voltage of the individual PV battery. Further, the ratio of P3 to P4 is determined as the required quantity N of the PV battery. Then, combining the battery status information set, N PV batteries are selected from the PV battery array as the target PV battery set. Specifically, photovoltaic (PV) batteries typically use lithium or lead-acid batteries. Excessive charging and discharging can affect battery life. Therefore, when PV batteries are needed to stabilize the power of load equipment, healthy PV batteries with relatively sufficient charge are prioritized to avoid overcharging and over-discharging of batteries with relatively low charge, thereby improving the overall battery life of the PV array. In particular, the required quantity N is determined based on the connection relationship of the PV batteries in the array (e.g., series connection, partial parallel connection with global series connection, partial series connection with global parallel connection) and the electrical relationship between current and voltage. Furthermore, to ensure redundancy, the value of N can be increased by adding a redundancy factor (e.g., 2) to the calculated result to avoid insufficient power when the generated power P1 decays below α.

[0092] Step S3: Generate a delay trigger based on the predicted occlusion start time included in the above occlusion description information.

[0093] The delay trigger is the start time used to trigger the target photovoltaic cell array to supply power according to the discharge power P3 and the photovoltaic power generation array according to the power generation power α×P1.

[0094] In practice, a delay trigger can be constructed by subtracting a preset duration (e.g., 1 second) from the predicted occlusion start time. This ensures the early switching of power supply equipment, thereby further guaranteeing power supply stability.

[0095] Step S4: In response to the triggering of the above-mentioned delay trigger, the power load corresponding to the above-mentioned load description information is supplied with integrated power according to the above-mentioned target photovoltaic cell set and photovoltaic power generation panel array.

[0096] In practice, the aforementioned implementing entity can control the target photovoltaic cell array to discharge according to the discharge power P3, and at the same time continue to generate electricity through the photovoltaic panel array, so as to achieve dual power supply for the load equipment.

[0097] Step 105: In response to the shading description information not meeting the triggering condition, the photovoltaic batteries in the photovoltaic battery array are controlled to store electricity according to the load description information.

[0098] In some embodiments, in response to the shading description information not meeting the triggering condition, the aforementioned execution entity performs energy storage control on the photovoltaic batteries in the photovoltaic battery array based on the load description information.

[0099] In practice, when the shading description information does not meet the triggering conditions, it indicates that there is no cumulus cloud shading the photovoltaic array. At this time, when the load power (P2) represented by the load description information is less than the power generation power (P1) of the photovoltaic array, the photovoltaic array can be used to store electricity for the photovoltaic batteries in the photovoltaic battery array.

[0100] In some optional implementations of certain embodiments, the execution entity, in response to the shading description information not meeting the triggering condition, performs energy storage control on the photovoltaic batteries in the photovoltaic battery array according to the load description information, including: Step S1: Determine the real-time power generation corresponding to the photovoltaic panel array.

[0101] In practice, real-time power generation can be read through a photovoltaic panel array monitoring system. This monitoring system is specifically designed to monitor the operational status of photovoltaic panels.

[0102] Step S2: Generate a power difference based on the real-time power generation and the load description information described above.

[0103] Wherein, power difference = real-time power generation - load power included in load description information.

[0104] Step S3: In response to the power difference being greater than the preset power threshold, determine the battery status of the photovoltaic battery in the photovoltaic battery array and obtain a set of battery status information.

[0105] Among them, battery status information represents the battery status of photovoltaic batteries. Battery status information may include, but is not limited to: battery identification, battery voltage, battery current, battery capacity, and battery temperature.

[0106] In practice, the battery status of each photovoltaic cell in a photovoltaic battery array can be read through a battery management system (BMS) to obtain the corresponding battery status information. The BMS is a system used to monitor the battery status of photovoltaic cells.

[0107] Step S4: Determine the average battery capacity based on the above set of battery status information.

[0108] Wherein, average battery capacity = sum(battery capacity included in the battery capacity information set) / len(battery capacity information set). sum() is the summation function. len() is the function that determines the length of the set.

[0109] Step S5: Select photovoltaic batteries from the above photovoltaic battery array whose current battery capacity is less than the above average battery capacity, and use them as photovoltaic batteries to be stored, thus obtaining a set of photovoltaic batteries to be stored.

[0110] Step S6: Using the power difference mentioned above as the energy storage power, perform battery energy storage on the photovoltaic battery array to be stored.

[0111] In practice, the aforementioned implementing entity can use the power difference as the storage power to perform parallel battery storage for the photovoltaic batteries in the array to be stored. Specifically, this method avoids overcharging photovoltaic batteries with higher charge levels and ensures timely replenishment of photovoltaic batteries with lower charge levels. Simultaneously, using the average battery charge level as a boundary can also relatively ensure that the photovoltaic batteries in the array are at approximately the same charge level.

[0112] In some optional implementations of some embodiments, the above method further includes: Step S1: Simultaneously display the battery status information set, the above-mentioned load description information, and the photovoltaic panel status information corresponding to the photovoltaic panel array on the remote control terminal.

[0113] Among them, the remote control terminal is a monitoring terminal used for remotely monitoring photovoltaic panel arrays, photovoltaic battery arrays and load equipment.

[0114] Step S2: Based on the direction of power flow, the corresponding twin power consumption model is synchronously updated on the remote control terminal.

[0115] Among them, the twin power consumption model is a twin digital model constructed based on the circuit connection relationship between the photovoltaic panel array, the photovoltaic battery array, and the load equipment.

[0116] In practice, depending on different power supply scenarios such as photovoltaic battery charging or discharging, and photovoltaic panel arrays supplying power independently, the current flow direction in the twin power consumption model can be updated in real time, thereby intuitively displaying the power generation and current usage.

[0117] The above embodiments of this disclosure have the following beneficial effects: The photovoltaic battery control method based on environmental perception, as described in some embodiments of this disclosure, reduces equipment costs and, through effective battery charging and discharging control, reduces the rate of battery degradation. Specifically, firstly, a sequence of sky images is acquired, where the sky images correspond to the sky above a target area, which is the area where the photovoltaic panel array is located. In practice, cumulus clouds are one of the main causes of short-term power generation changes; therefore, sky images are acquired for subsequent cumulus cloud identification. Secondly, based on the wind state information corresponding to the target area, cumulus cloud identification is performed on each sky image in the above sky image sequence to obtain a cumulus cloud description information sequence, where the cumulus cloud description information includes: cumulus cloud location, cumulus cloud movement speed, and cumulus cloud confidence level. Through cumulus cloud identification, the distribution and changes of cumulus clouds above the photovoltaic panel array are obtained. Next, based on the above cumulus cloud description information sequence and solar position information, shading description information is determined, where the shading description information includes: predicted shading start time, predicted shading duration, and shading confidence level. In practice, when cumulus clouds are located between the sun and photovoltaic (PV) panels, they can affect the instantaneous power generation of the PV panels. Therefore, this disclosure combines cumulus cloud description information sequences and solar position information to determine the start time, duration, and corresponding confidence level of shading through shading prediction. Furthermore, in response to the aforementioned shading description information meeting the triggering conditions, the PV batteries in the PV battery array are discharged based on the aforementioned shading description information and load description information, wherein the PV batteries store electricity through the PV panel array. By combining the shading description information and load description information, the discharge control of the PV batteries is achieved, thereby reducing the impact of cumulus clouds on the instantaneous power generation of PV power, thus ensuring the stable operation of the load equipment. Finally, in response to the aforementioned shading description information not meeting the triggering conditions, the PV batteries in the aforementioned PV battery array are charged based on the aforementioned load description information. This achieves intelligent charging control of PV batteries while ensuring the stable operation of the load equipment. This method eliminates the need for a large number of inverters and other power equipment, thereby reducing equipment costs, and effectively controlling battery charging and discharging reduces the rate of battery degradation.

[0118] Further reference Figure 5 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a photovoltaic battery control device based on environmental perception. These device embodiments are similar to... Figure 1Corresponding to the method embodiments shown, this environment-aware photovoltaic battery control device can be specifically applied to various electronic devices.

[0119] like Figure 5 As shown, some embodiments of the photovoltaic battery control device 500 based on environmental perception include: an acquisition unit 501, a cumulus cloud recognition unit 502, a determination unit 503, a discharge control unit 504, and a battery storage control unit 505. The acquisition unit 501 is configured to acquire a sequence of sky images, wherein the sky images correspond to the sky above a target area, which is an area where a photovoltaic panel array is installed. The cumulus cloud recognition unit 502 is configured to perform cumulus cloud recognition on each sky image in the sky image sequence based on wind state information corresponding to the target area, obtaining a sequence of cumulus cloud description information, wherein the cumulus cloud description information includes: cumulus cloud location, cumulus cloud movement speed, and cumulus cloud confidence level. The determination unit... 503 is configured to determine shading description information based on the above-mentioned cumulus description information sequence and solar position information, wherein the shading description information includes: predicted shading start time, predicted shading duration, and shading confidence level; discharge control unit 504 is configured to, in response to the above-mentioned shading description information meeting the trigger condition, perform discharge control on the photovoltaic batteries in the photovoltaic battery array based on the above-mentioned shading description information and load description information, wherein the photovoltaic batteries store electricity through the photovoltaic panel array; energy storage control unit 505 is configured to, in response to the above-mentioned shading description information not meeting the above-mentioned trigger condition, perform energy storage control on the photovoltaic batteries in the photovoltaic battery array based on the above-mentioned load description information.

[0120] It is understandable that the units described in the environmentally sensitive photovoltaic battery control device 500 are related to the reference... Figure 1 The steps in the described method correspond accordingly. Therefore, the operations, features, and beneficial effects described above for the method also apply to the environmentally aware photovoltaic battery control device 500 and the units contained therein, and will not be repeated here.

[0121] The following is for reference. Figure 6 It shows a schematic diagram of the structure of an electronic device (e.g., a computing device) 600 suitable for implementing some embodiments of the present disclosure. Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0122] like Figure 6As shown, the electronic device 600 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory 602 or a program loaded from a storage device 608 into a random access memory 603. The random access memory 603 also stores various programs and data required for the operation of the electronic device 600. The processing unit 601, the read-only memory 602, and the random access memory 603 are interconnected via a bus 604. An input / output interface 605 is also connected to the bus 604.

[0123] Typically, the following devices can be connected to the input / output interface 605: input devices 606 including, for example, a touchscreen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 607 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 608 including, for example, magnetic tape, hard disk, etc.; and communication devices 609. Communication device 609 allows electronic device 600 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 An electronic device 600 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 6 Each box shown can represent a device or multiple devices as needed.

[0124] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a storage device 608, or installed from a read-only memory 602. When the computer program is executed by the processing device 601, it performs the functions defined above in the methods of some embodiments of this disclosure.

[0125] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0126] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0127] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire a sequence of sky images, wherein the sky images correspond to the sky above a target area, the target area being an area where a photovoltaic panel array is installed; identify cumulus clouds in each sky image in the sky image sequence based on wind state information corresponding to the target area, obtaining a sequence of cumulus cloud description information, wherein the cumulus cloud description information includes: cumulus cloud location, cumulus cloud movement speed, and cumulus cloud confidence; determine shading description information based on the cumulus cloud description information sequence and solar position information, wherein the shading description information includes: predicted shading start time, predicted shading duration, and shading confidence; in response to the shading description information satisfying a trigger condition, perform discharge control on the photovoltaic batteries in the photovoltaic battery array based on the shading description information and load description information, wherein the photovoltaic batteries store electricity through the photovoltaic panel array; in response to the shading description information not satisfying the trigger condition, perform energy storage control on the photovoltaic batteries in the photovoltaic battery array based on the load description information.

[0128] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0129] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0130] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0131] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for controlling a photovoltaic battery based on environmental perception, characterized in that, include: Acquire a sequence of sky images, wherein the sky images correspond to the sky above a target area, and the target area is the area where a photovoltaic panel array is set up; Based on the wind state information corresponding to the target area, cumulus cloud identification is performed on each sky image in the sky image sequence to obtain a cumulus cloud description information sequence, wherein the cumulus cloud description information includes: cumulus cloud location, cumulus cloud movement speed, and cumulus cloud confidence level; Based on the cumulus cloud description information sequence and the sun position information, occlusion description information is determined, wherein the occlusion description information includes: predicted occlusion start time, predicted occlusion duration, and occlusion confidence level; In response to the shading description information satisfying the triggering condition, the photovoltaic batteries in the photovoltaic battery array are discharged according to the shading description information and the load description information, wherein the photovoltaic batteries store electricity through the photovoltaic panel array; In response to the shading description information not meeting the trigger condition, the photovoltaic batteries in the photovoltaic battery array are controlled to store electricity according to the load description information.

2. The photovoltaic battery control method based on environmental perception according to claim 1, characterized in that, The step of responding to the shading description information satisfying the trigger condition, and controlling the discharge of the photovoltaic cells in the photovoltaic battery array according to the shading description information and the load description information, includes: Determine the battery status of the photovoltaic batteries in the photovoltaic battery array to obtain a set of battery status information; Based on the load description information, the battery status information set, the shading description information including the predicted shading duration and the preset power ratio, a target photovoltaic battery set is determined, wherein the target photovoltaic battery is the photovoltaic battery in the photovoltaic battery array used to supply power to the application electrical load according to the load description information; A delay trigger is generated based on the predicted occlusion start time included in the occlusion description information; In response to the triggering of the delay trigger, the power load corresponding to the load description information is supplied in a fusion manner according to the target photovoltaic cell set and photovoltaic power generation panel array.

3. The photovoltaic battery control method based on environmental perception according to claim 1, characterized in that, In response to the shading description information not meeting the trigger condition, the photovoltaic battery in the photovoltaic battery array is controlled for energy storage based on the load description information, including: Determine the real-time power generation corresponding to the photovoltaic panel array; A power difference is generated based on the real-time power generation and the load description information; In response to the power difference being greater than a preset power threshold, the battery status of the photovoltaic battery in the photovoltaic battery array is determined, and a set of battery status information is obtained; The average battery capacity is determined based on the battery status information set. From the photovoltaic battery array, select the photovoltaic batteries whose current battery capacity is less than the average battery capacity, and use them as photovoltaic batteries to be stored, thus obtaining a set of photovoltaic batteries to be stored. The power difference is used as the storage power to store electricity in the photovoltaic battery array to be stored.

4. The photovoltaic battery control method based on environmental perception according to claim 2 or 3, characterized in that, The method further includes: The remote control terminal simultaneously displays the battery status information set, the load description information, and the photovoltaic panel status information corresponding to the photovoltaic panel array. Based on the direction of power flow, the corresponding twin power consumption model is synchronously updated on the remote control terminal.

5. The photovoltaic battery control method based on environmental perception according to claim 4, characterized in that, The step involves identifying cumulus clouds in each sky image of the sky image sequence based on the wind state information corresponding to the target area, to obtain a sequence of cumulus cloud description information, including: The sky image is converted to a different color space to obtain the converted sky image. Image enhancement is performed on the converted sky image to obtain the enhanced sky image; By using a pre-trained cumulus recognition model, cumulus recognition is performed on the enhanced sky image to obtain cumulus description information corresponding to the sky image, including cumulus location and cumulus confidence level. In response to the fact that the sky image is the first sky image in the sky image sequence, the preset speed is determined to be the cumulus movement speed included in the cumulus description information corresponding to the sky image; In response to the fact that the sky image is not the first sky image in the sky image sequence, the cumulus movement speed included in the cumulus description information corresponding to the sky image is generated based on the previous cumulus position and the cumulus position included in the cumulus description information corresponding to the sky image.

6. The photovoltaic battery control method based on environmental perception according to claim 5, characterized in that, The cumulus cloud recognition model includes: a backbone feature extraction module, a multi-scale feature extraction module, a morphological feature recognition module, and a detection head group. The morphological feature recognition module includes: a cumulus cloud edge feature extractor, a cumulus cloud texture feature extractor, and a cumulus cloud color feature extractor. The detection head group consists of a first detection head, a second detection head, and a third detection head. The first detection head is used for cumulus cloud boundary point localization, the second detection head is used for cumulus cloud center point localization, and the third detection head is used for cumulus cloud confidence generation. The model, through pre-trained training, performs cumulus cloud recognition on the enhanced sky image to obtain cumulus cloud description information corresponding to the sky image, including cumulus cloud location and cumulus cloud confidence, including: The enhanced sky image is coarsely extracted using the backbone feature extraction module to obtain the initial sky image features. Multi-scale feature extraction is performed based on the multi-scale feature extraction module and the initial sky image features to obtain a set of multi-scale sky image feature groups. The multi-scale sky image feature groups consist of a first multi-scale sky image feature and a second multi-scale sky image feature. The first multi-scale sky image feature is extracted by a horizontal convolution operator, and the second multi-scale sky image feature is extracted by a vertical convolution operator. Based on the wind state information, determine the horizontal and vertical weights; For each multi-scale sky image feature group in the multi-scale sky image feature group set, the first multi-scale sky image features and the second multi-scale sky image features included in the multi-scale sky image feature group are superimposed with feature weights according to the horizontal weights and the vertical weights to obtain fused image features; Cumulus cloud morphology is identified based on the morphological feature recognition module and the obtained fused image feature set to obtain cumulus cloud morphology features, wherein the cumulus cloud morphology features are composed of cumulus cloud edge features, cumulus cloud texture features and cumulus cloud color features. Based on the detection head group and the cumulus morphology features, cumulus description information corresponding to the sky image is generated, including the cumulus location and cumulus confidence level.

7. The photovoltaic battery control method based on environmental perception according to claim 6, characterized in that, The step of determining the occlusion description information based on the cumulus cloud description information sequence and the sun position information includes: Based on the solar position information and the area position information corresponding to the target area, a fence space is generated, wherein the fence space is a quadrangular pyramidal space; Temporal location features are extracted from the cumulus location information included in the cumulus description information sequence to generate cumulus location features; Temporal motion features are extracted from the cumulus movement speed included in the cumulus description information sequence to generate cumulus movement speed features; Based on the cumulus location characteristics and the cumulus movement speed characteristics, the predicted location and prediction confidence level are determined; When the predicted location is within the fence space, the following processing steps are performed: The average cumulus movement speed is determined based on the cumulus movement speed included in the cumulus description information sequence. Based on the average cumulus movement speed, the predicted location, the fence space, and the target cumulus location, the predicted occlusion start time and predicted occlusion duration included in the occlusion description information are determined, wherein the target cumulus location is the cumulus location included in the cumulus description information located at the end of the sequence in the cumulus description information sequence; The prediction confidence is determined as the occlusion confidence included in the occlusion description information.

8. A photovoltaic battery control device based on environmental perception, characterized in that, include: The acquisition unit is configured to acquire a sequence of sky images, wherein the sky images correspond to the sky above a target area, and the target area is the area where a photovoltaic panel array is set up; The cumulus cloud recognition unit is configured to perform cumulus cloud recognition on each sky image in the sky image sequence based on the wind state information corresponding to the target area, and obtain a cumulus cloud description information sequence, wherein the cumulus cloud description information includes: cumulus cloud location, cumulus cloud movement speed and cumulus cloud confidence level; The determining unit is configured to determine occlusion description information based on the cumulus description information sequence and the sun position information, wherein the occlusion description information includes: predicted occlusion start time, predicted occlusion duration, and occlusion confidence. A discharge control unit is configured to control the discharge of photovoltaic batteries in a photovoltaic battery array in response to the shading description information meeting a trigger condition, based on the shading description information and load description information, wherein the photovoltaic batteries store electricity through a photovoltaic panel array. The energy storage control unit is configured to control the energy storage of the photovoltaic cells in the photovoltaic battery array according to the load description information in response to the shading description information not meeting the trigger condition.

9. An electronic device, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 7.

10. A computer-readable medium, characterized in that, It stores a computer program thereon, wherein the computer program, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images