Charging pile fire monitoring method, electronic device and storage medium

Abstract

This application discloses a charging pile fire monitoring method, electronic device, and storage medium. The charging pile fire monitoring method includes: obtaining a current charging pile image of the charging pile; inputting the current charging pile image into a target feature extraction network model to obtain a target feature map, wherein the target feature extraction network model is based on a ResNet50 network model and a semantic mining model, and the semantic mining model is used to extract and associate the global semantic features of the current charging pile image; and inputting the target feature map into a classifier to obtain a fire monitoring result. This method implements charging pile fire monitoring using a target feature extraction network model and classifier constructed based on the ResNet50 network model and semantic mining model. This can avoid missing charging pile fires due to manual inspections of charging piles, thereby improving the accuracy of fire monitoring for charging piles.

Description

Technical Field

[0001] The present application belongs to the field of image processing technology, and in particular relates to a fire monitoring method, electronic equipment, and storage medium for a charging pile. Background Art

[0002] With the development of electric vehicles, electric vehicles and electric bicycles occupy a large proportion in people's daily lives. Due to the characteristics of electric vehicles, electric vehicles need to be charged frequently, so charging piles have emerged.

[0003] Charging piles are mostly installed in open-air stations, where high temperatures are unavoidable. This harsh environment, coupled with the long-term high current operation of battery components, leads to relatively serious aging and damage. This is especially true during the hot summer months, when surface temperatures can reach over 40°C. These factors cause battery components to heat up rapidly, posing a significant risk of electrical fires. Therefore, fire monitoring of charging piles is urgently needed.

[0004] Currently, fire monitoring of charging piles relies primarily on manual inspections to detect fires. However, during manual inspections, inspectors can easily miss fires due to visual fatigue, resulting in low accuracy in fire monitoring of charging piles. Summary of the Invention

[0005] In view of this, embodiments of the present application provide a fire monitoring method, electronic device, and storage medium for a charging pile to overcome the above problems of the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for monitoring a fire in a charging pile, comprising:

[0007] Get the current charging pile image of the charging pile;

[0008] Inputting the current charging pile image into a target feature extraction network model to obtain a target feature map, wherein the target feature extraction network model is obtained based on a ResNet50 network model and a semantic mining model, and the semantic mining model is used to extract and associate global semantic features of the current charging pile image;

[0009] The target feature map is input into the classifier to obtain the fire monitoring result.

[0010] In some optional embodiments, before inputting the target feature map into the classifier to obtain the fire monitoring result, the charging pile fire monitoring method further includes:

[0011] Performing feature aggregation on the target feature map to obtain an aggregated feature map;

[0012] The target feature map is input to the classifier to obtain a fire monitoring result;

[0013] The aggregated feature map is input into the classifier to obtain the fire monitoring result.

[0014] In some optional embodiments, performing feature aggregation on the target feature map to obtain an aggregated feature map includes:

[0015] Performing global average pooling and global maximum pooling on the target feature map respectively to obtain corresponding global average pooling feature map and global maximum pooling feature map;

[0016] The global average pooling feature map and the global maximum pooling feature map are fused to obtain the aggregated feature map.

[0017] In some optional embodiments, before obtaining the current charging pile image of the charging pile, the charging pile fire monitoring method further includes:

[0018] Constructing the semantic mining model;

[0019] The ResNet50 network model and the semantic mining model are integrated to obtain an initial feature extraction network model;

[0020] The initial feature extraction network model is trained to obtain the target feature extraction network model.

[0021] In some optional embodiments, the ResNet50 network model includes a two-dimensional convolution module, a first stage module, a second stage module, a third stage module, and a fourth stage module connected in sequence. The fusion of the ResNet50 network model and the semantic mining model to obtain an initial feature extraction network model includes:

[0022] The semantic mining model is spliced ​​between the third stage module and the fourth stage module of the ResNet50 network model to obtain the initial feature extraction network model.

[0023] In some optional embodiments, the constructing of the semantic mining model includes:

[0024] fusing a first transposition module, a flattening module, a layer normalization module, a linear layer encoding module, a linear layer decoding module, a rectified linear unit activation function module, a size conversion module, and a second transposition module to obtain the semantic mining model;

[0025] The first transposition module is used to transpose the received initial feature map from the first dimension to the second dimension to obtain a first feature map; the flattening module is used to flatten the dimension of the first feature map to obtain a second feature map; the layer normalization module is used to perform layer normalization processing on the second feature map to obtain a third feature map; the linear layer encoding module is used to perform linear transformation and mapping on the third feature map to obtain a fourth feature map; the linear layer decoding module is used to extract and associate the global features of the fourth feature map to obtain a fifth feature map; the rectified linear unit activation function module is used to perform rectified linear unit activation function processing on the fifth feature map to obtain a sixth feature map; the size conversion module is used to convert the dimension of the sixth feature map to the second dimension to obtain a seventh feature map; the second transposition module is used to transpose the seventh feature map from the second dimension to the first dimension to obtain and output an eighth feature map.

[0026] In some optional embodiments, obtaining the current charging pile image of the charging pile includes:

[0027] Performing short circuit detection and open circuit detection on the charging pile to obtain detection results;

[0028] When it is determined according to the detection result that the charging pile is in an available charging state, the camera is controlled to capture an image of the charging pile to obtain the current charging pile image.

[0029] In some optional embodiments, the charging pile fire monitoring method further includes:

[0030] When the fire monitoring result includes a first monitoring result, fire warning information is generated, and a fire extinguishing device is controlled to extinguish the fire on the charging pile. The first monitoring result is used to indicate that the current charging pile image contains flame information.

[0031] In a second aspect, an embodiment of the present application provides a fire monitoring device for a charging pile, and the fire monitoring device for the charging pile includes an acquisition module, a first input module, and a second input module.

[0032] The acquisition module is used to acquire the current charging pile image of the charging pile;

[0033] The first input module is used to input the current charging pile image into the target feature extraction network model to obtain a target feature map, wherein the target feature extraction network model is obtained based on the ResNet50 network model and the semantic mining model, and the semantic mining model is used to extract and associate the global semantic features of the current charging pile image;

[0034] The second input module is used to input the target feature map into the classifier to obtain the fire monitoring result.

[0035] In a third aspect, an embodiment of the present application provides an electronic device comprising a memory; one or more processors coupled to the memory; and one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to execute the fire monitoring method for a charging pile provided in the first aspect above.

[0036] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a program code is stored. The program code can be called by a processor to execute the fire monitoring method for a charging pile as provided in the first aspect above.

[0037] In a fifth aspect, an embodiment of the present application provides a computer program product, which, when running on a computer device, enables the computer device to execute the fire monitoring method for a charging pile provided in the first aspect above.

[0038] The solution provided in the present application obtains a current charging pile image of the charging pile and inputs the current charging pile image into a target feature extraction network model to obtain a target feature map. The target feature extraction network model is obtained based on the ResNet50 network model and the semantic mining model. The semantic mining model is used to extract and associate the global semantic features of the current charging pile image, and input the target feature map into a classifier to obtain a fire monitoring result. The solution realizes fire monitoring of charging piles based on the target feature extraction network model and classifier constructed based on the ResNet50 network model and the semantic mining model, which can avoid missing the flames of charging piles due to manual inspections of charging piles, thereby improving the monitoring accuracy of fire monitoring of charging piles.

[0039] The semantic mining model is used to extract and associate the global semantic features of the current charging pile image. A target feature extraction network model is constructed based on the ResNet50 network model and the semantic mining model, which improves the accuracy of the target feature extraction network model in extracting flame features, making small flames easier to detect, which is conducive to further improving the monitoring accuracy of fire monitoring of charging piles. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0041] Figure 1A schematic diagram of a scenario of a charging pile fire monitoring system provided in an embodiment of the present application is shown.

[0042] Figure 2 A flow chart of a fire monitoring method for a charging pile provided in an embodiment of the present application is shown.

[0043] Figure 3 A structural schematic diagram of a semantic mining model in the fire monitoring method for charging piles provided in an embodiment of the present application is shown.

[0044] Figure 4 Another flow chart of the fire monitoring method for a charging pile provided in an embodiment of the present application is shown.

[0045] Figure 5 A flow chart of feature aggregation in the fire monitoring method for a charging pile provided in an embodiment of the present application is shown.

[0046] Figure 6 Another flow chart of the fire monitoring method for a charging pile provided in an embodiment of the present application is shown.

[0047] Figure 7 A structural schematic diagram of an initial feature extraction network model in the fire monitoring method for a charging pile provided in an embodiment of the present application is shown.

[0048] Figure 8 A structural block diagram of a fire monitoring device for a charging pile provided in an embodiment of the present application is shown.

[0049] Figure 9 A functional block diagram of an electronic device provided in an embodiment of the present application is shown.

[0050] Figure 10 A computer-readable storage medium provided in an embodiment of the present application is shown for storing or carrying program code for implementing the fire monitoring method for a charging pile provided in an embodiment of the present application.

[0051] Figure 11 A computer program product provided in an embodiment of the present application is shown for storing or carrying program code for implementing the fire monitoring method for a charging pile provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the purpose, features, and advantages of the invention of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the embodiments described below are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0053] It will be understood that when used in this specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0054] It should also be understood that the terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0055] It should be further understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0056] In addition, in the description of the present application, the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0057] With the development of electric vehicles, electric vehicles and electric bicycles occupy a large proportion in people's daily lives. Due to the characteristics of electric vehicles, electric vehicles need to be charged frequently, so charging piles have emerged.

[0058] Charging piles are mostly installed in open-air stations, where high temperatures are unavoidable. This harsh environment, coupled with the long-term high current operation of battery components, leads to relatively serious aging and damage. This is especially true during the hot summer months, when surface temperatures can reach over 40°C. These factors cause battery components to heat up rapidly, posing a significant risk of electrical fires. Therefore, fire monitoring of charging piles is urgently needed.

[0059] Currently, fire monitoring of charging piles relies primarily on manual inspections to detect fires. However, during manual inspections, inspectors can easily miss fires due to visual fatigue, resulting in low accuracy in fire monitoring of charging piles.

[0060] In response to the above problems, the embodiments of the present application provide a fire monitoring method, electronic device and storage medium for charging piles, which obtain a current charging pile image of the charging pile and input the current charging pile image into a target feature extraction network model to obtain a target feature map. The target feature extraction network model is obtained based on the ResNet50 network model and the semantic mining model. The semantic mining model is used to extract and associate the global semantic features of the current charging pile image, and input the target feature map into a classifier to obtain fire monitoring results. The target feature extraction network model and classifier constructed based on the ResNet50 network model and the semantic mining model are used to perform fire monitoring on the charging pile, which can avoid missing the flames of the charging pile due to manual inspection of the charging pile, and improve the monitoring accuracy of fire monitoring on the charging pile.

[0061] The semantic mining model is used to extract and associate the global semantic features of the current charging pile image. A target feature extraction network model is constructed based on the ResNet50 network model and the semantic mining model, which improves the accuracy of the target feature extraction network model in extracting flame features, making small flames easier to detect, which is conducive to further improving the monitoring accuracy of fire monitoring of charging piles.

[0062] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application.

[0063] See also Figure 1 , which shows a schematic diagram of an application scenario of the charging pile fire monitoring system provided in an embodiment of the present application. The charging pile fire monitoring system may include a charging pile 100, a camera 200 and a processing device 300. The camera 200 can be connected to the processing device 300 through a network and exchange data with the processing device 300 through the network.

[0064] The charging pile 100 may include an alternating current (AC) charging pile or a direct current (DC) charging pile, and is not limited here.

[0065] The camera 200 can be used to capture images of the charging pile 100 , obtain a charging pile image, and send the charging pile image to the processing device 300 via a network.

[0066] The camera 200 may include any one of a wide-angle camera, a macro camera, an ultra-wide-angle camera, a panoramic camera, a depth camera, a monocular camera, and a binocular camera. The type of the camera 200 is not limited here and can be set according to actual needs.

[0067] The processing device 300 can be used to receive the charging pile image sent by the camera 200 and perform fire monitoring on the charging pile based on the charging pile image.

[0068] The processing device 300 can be any one of a server and a terminal device, etc., which is not limited here and can be specifically configured according to actual needs.

[0069] The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), big data, and artificial intelligence platforms.

[0070] The terminal device may be a mobile terminal device (eg, a vehicle-mounted terminal, a PDA (Personal Digital Assistant), a tablet PC (Tablet Personal Computer, Tablet PC), a laptop computer, etc.), or a fixed terminal device (eg, a desktop computer, a smart panel, etc.).

[0071] The network may be any one of a ZigBee network, a Bluetooth (BT) network, a Wireless Fidelity (Wi-Fi) network, a Thread network, a Long Range Radio (LoRa) network, a Low-Power Wide-Area Network (LPWAN), an infrared network, a Narrow Band Internet of Things (NB-IoT), a Controller Area Network (CAN), a Digital Living Network Alliance (DLNA) network, a Wide Area Network (WAN), a Local Area Network (LAN), a Metropolitan Area Network (MAN), or a Wireless Personal Area Network (WPAN), etc., without limitation herein.

[0072] See also Figure 2, which shows a flow chart of a fire monitoring method for a charging pile provided by an embodiment of the present application. In a specific embodiment, the fire monitoring method for a charging pile can be applied to Figure 1 The processing device 300 in the charging pile fire monitoring system shown in FIG. 3 is used as an example to Figure 2 The process shown in FIG. 1 is described in detail. The fire monitoring method for a charging pile may include the following steps 110 to 130.

[0073] Step 110: Acquire the current charging pile image of the charging pile.

[0074] In an embodiment of the present application, when the inspector needs to perform fire monitoring on the charging pile, he can send a monitoring instruction to the processing device, the processing device receives and responds to the monitoring instruction, sends an image acquisition instruction to the camera via the network, the camera receives and responds to the image acquisition instruction, performs image acquisition on the charging pile, obtains the current charging pile image, and sends the current charging pile image to the processing device via the network, and the processing device receives the current charging pile image returned by the camera.

[0075] In some embodiments, the charging pile fire monitoring system may also include a battery management system (BMS). The BMS may be located relative to the charging pile and configured to detect short circuits and / or open circuits on the charging pile. The BMS is connected to the processing device via a network and exchanges data with the processing device via the network.

[0076] When the inspector needs to monitor the charging pile for fire, he can send a monitoring instruction to the processing device. The processing device receives and responds to the monitoring instruction and sends a detection instruction to the BMS through the network. The BMS receives and responds to the detection instruction, performs short circuit detection and open circuit detection on the charging pile, obtains the detection result, and sends the detection result to the processing device through the network. The processing device receives the detection result returned by the BMS, determines whether the charging pile is in a usable charging state based on the detection result, and when it is determined that the charging pile is in a usable charging state based on the detection result, sends an image acquisition instruction to the camera through the network. The camera receives and responds to the image acquisition instruction, performs image acquisition on the charging pile, obtains the current charging pile image, and sends the current charging pile image to the processing device through the network. The processing device receives the current charging pile image returned by the camera.

[0077] The detection result may include a first detection result for characterizing that the charging pile is in a short-circuit state or an open-circuit state, and a second detection result for characterizing that the charging pile is in a non-short-circuit state and a non-short-circuit state.

[0078] When the detection result is the first detection result, it is determined that the charging pile is in an unavailable charging state; when the detection result is the second detection result, it is determined that the charging pile is in an available charging state.

[0079] The BMS can control the charging pile to start working, collect the output current of the charging pile to the load connected to the charging pile, obtain the output current, and determine the detection result based on the output current.

[0080] When the output current is 0 (indicating that the charging pile is in an open circuit state), or the output current is greater than or equal to the current threshold (indicating that the charging pile is in a short circuit state), the detection result is determined to be the first detection result; when the output current is greater than 0 and less than the current threshold, the detection result is determined to be the second detection result.

[0081] The current threshold can be used to characterize the minimum current when the charging pile is in a short-circuit state. The current threshold can be a current pre-set by an inspector, or a current automatically generated by a processing device based on multiple fire monitoring processes of the charging pile, etc., and is not limited here.

[0082] In some embodiments, the processing device can detect the inspector's operation. When it is determined based on the detected inspector's operation that the inspector has input a monitoring instruction for fire monitoring of the charging pile, the monitoring instruction for fire monitoring of the charging pile is received.

[0083] For example, when an inspector needs to monitor a charging pile for fire, he or she can perform touch operations on the operation panel of the processing device. The processing device responds to the inspector's touch operations, generates corresponding touch signals, and analyzes the touch signals. When it is determined that the touch signal is a preset monitoring signal for representing fire monitoring of the charging pile, it is determined that a monitoring instruction for fire monitoring of the charging pile has been received.

[0084] In some embodiments, the processing device may be provided with a voice recognition module. When the inspector needs to perform fire monitoring on the charging pile, the inspector may send voice information within the voice collection range of the voice recognition module. The voice recognition module collects the voice information sent by the inspector and performs voice recognition on the collected voice information. Based on the recognition results of the voice recognition, it is determined that the recognition results contain keywords for indicating fire monitoring of the charging pile, such as "charging pile fire monitoring", or "charging pile" and "fire monitoring", etc., and it is determined that a monitoring instruction for fire monitoring of the charging pile has been received.

[0085] As an example, the voice message sent by the inspector is: fire monitoring of charging piles, and the recognition result of the voice recognition contains the keywords "charging pile" and "fire monitoring", then it is determined that the monitoring instruction for fire monitoring of charging piles has been received.

[0086] In some embodiments, the charging pile fire monitoring system may further include a client associated with an inspector, where the client is connected to the processing device via a network and exchanges data with the processing device via the network.

[0087] When the inspector needs to monitor the charging pile for fire, he can send a monitoring instruction to the client. The client receives and responds to the monitoring instruction, forwards the monitoring instruction to the processing device through the network, and the processing device receives the monitoring instruction forwarded by the client.

[0088] Among them, the client can be any one of a mobile client (for example, a mobile phone client, a PDA client, a Tablet PC client, a laptop client, a smart watch client, a smart bracelet client or a wearable client, etc.) or a fixed client (for example, a desktop computer client, a smart panel client, etc.). The type of client is not limited here and can be set according to actual needs.

[0089] Step 120: Input the current charging pile image into the target feature extraction network model to obtain a target feature map.

[0090] In an embodiment of the present application, after the processing device obtains the current charging pile image of the charging pile, the current charging pile image can be input into the target feature extraction network model. The target feature extraction network model receives and responds to the current charging pile image, and outputs the target feature map to the processing device. The processing device receives the target feature map output by the target feature extraction network model.

[0091] Among them, the target feature extraction network model can be obtained based on the ResNet50 network model and the semantic mining model, and the semantic mining model can be used to extract and associate the global semantic features of the current charging pile image.

[0092] The ResNet50 network model may include a two-dimensional convolution module, a first stage module, a second stage module, a third stage module, and a fourth stage module connected in sequence.

[0093] The first stage module may include three bottleneck layers, the second stage 2 may include four bottleneck layers, the third stage 3 may include six bottleneck layers, and the fourth stage 4 may include three bottleneck layers. The specific structure of the ResNet50 network model and its bottleneck layer structure are already known and will not be repeated here.

[0094] The semantic mining model may include a first transposition (first Permute function) module, a flattening (Flatten function) module, a layer normalization (LayerNom) module, a linear layer encoding (Linear Encoder) module, a linear layer decoding (Linear Decoder) module, a rectified linear unit activation (Rectified Linear Units, ReLU) function module, a size conversion (Reshape) module, and a second transposition (second Permute function) module connected in sequence, such as Figure 3 shown.

[0095] The first Permute function module can be used to transpose the received initial feature map from the first dimension to the second dimension to obtain the first feature map; the Flatten function module can be used to flatten the dimension of the first feature map to obtain the second feature map; the LayerNom module can be used to perform layer normalization on the second feature map to obtain the third feature map; the LinearEncoder module can be used to perform linear transformation and mapping on the third feature map to obtain the fourth feature map; the Linear Decoder module can be used to extract and associate the global features of the fourth feature map to obtain the fifth feature map; the ReLU function module can be used to perform ReLU function processing on the fifth feature map to obtain the sixth feature map; the Reshape module can be used to convert the dimension of the sixth feature map to the second dimension to obtain the seventh feature map; the second Permute function module can be used to transpose the seventh feature map from the second dimension to the first dimension to obtain and output the eighth feature map.

[0096] In some embodiments, after the processing device obtains the current charging pile image of the charging pile, it can preprocess the current charging pile image to obtain multiple preprocessed images, and input the multiple preprocessed images into the target feature extraction network model. The target feature extraction network model receives and responds to the multiple preprocessed images, and outputs the target feature map to the processing device. The processing device receives the target feature map output by the target feature extraction network model and preprocesses the current charging pile image, which can increase the feature richness of the charging pile image and help improve the extraction accuracy of the target feature map by the target feature extraction network model.

[0097] The preprocessing may include at least one of horizontal flipping, random left / right rotation at a preset angle, random preset ratio cropping, and window scaling to a preset window size, etc., which are not limited here.

[0098] As an example, the preset angle may be 1°-10°, the preset ratio may be 1 / 2 of the initial window or 1 / 4 of the initial window, and the preset window size may be 640×480, which is not limited here.

[0099] Step 130: Input the target feature map into the classifier to obtain the fire monitoring result.

[0100] In an embodiment of the present application, the processing device inputs the current charging pile image into the target feature extraction network model. After obtaining the target feature map, the target feature map can be input into the classifier. The classifier receives and responds to the target feature map and outputs the fire monitoring result to the processing device. The processing device receives the fire monitoring result output by the classifier, and realizes fire monitoring of the charging pile based on the target feature extraction network model and classifier constructed based on the ResNet50 network model and the semantic mining model. It can avoid the missed detection of the flame of the charging pile due to manual inspection of the charging pile, and improves the monitoring accuracy of the fire monitoring of the charging pile.

[0101] The semantic mining model is used to extract and associate the global semantic features of the current charging pile image. A target feature extraction network model is constructed based on the ResNet50 network model and the semantic mining model, which improves the accuracy of the target feature extraction network model in extracting flame features, making small flames easier to detect, which is conducive to further improving the monitoring accuracy of fire monitoring of charging piles.

[0102] The fire monitoring result may include a first monitoring result used to characterize that the current charging pile image contains flame information, and a second monitoring result used to characterize that the current charging pile image does not contain flame information.

[0103] The classifier may include any one of a Softmax classifier, a multiclass support vector machine (Multiclass Support Vector Machine, Multiclass SVM) classifier, a perceptron classifier, and a logistic regression classifier, etc., and is not limited here.

[0104] In some embodiments, the charging pile fire monitoring system may further include a fire extinguishing device, which may be positioned relative to the charging pile and used to extinguish a fire on the charging pile. The fire extinguishing device may be connected to the processing device via a network and exchange data with the processing device via the network.

[0105] When the processing device inputs the target feature map into the classifier and obtains the first monitoring result, it can generate fire warning information and send a fire extinguishing instruction to the fire extinguishing equipment through the network. The fire extinguishing equipment receives and responds to the fire extinguishing instruction to extinguish the fire of the charging pile. This realizes timely control of the fire extinguishing equipment to extinguish the fire of the charging pile in the early stage of monitoring the fire in the charging pile, which is conducive to reducing the losses caused by the fire in the charging pile.

[0106] The fire extinguishing equipment may include at least any one of water spraying fire extinguishing equipment, dry powder fire extinguishing equipment, foam fire extinguishing equipment, and gas fire extinguishing equipment, etc., which is not limited here.

[0107] The fire warning information may include at least any one of text warning information, sound warning information, and light warning information, etc., which is not limited here.

[0108] The solution provided in the present application obtains a current charging pile image of the charging pile and inputs the current charging pile image into a target feature extraction network model to obtain a target feature map. The target feature extraction network model is obtained based on the ResNet50 network model and the semantic mining model. The semantic mining model is used to extract and associate the global semantic features of the current charging pile image, and input the target feature map into a classifier to obtain a fire monitoring result. The solution realizes fire monitoring of charging piles based on the target feature extraction network model and classifier constructed based on the ResNet50 network model and the semantic mining model, which can avoid missing the flames of charging piles due to manual inspections of charging piles, thereby improving the monitoring accuracy of fire monitoring of charging piles.

[0109] The semantic mining model is used to extract and associate the global semantic features of the current charging pile image. A target feature extraction network model is constructed based on the ResNet50 network model and the semantic mining model, which improves the accuracy of the target feature extraction network model in extracting flame features, making small flames easier to detect, which is conducive to further improving the monitoring accuracy of fire monitoring of charging piles.

[0110] See also Figure 4 , which shows a flow chart of a fire monitoring method for a charging pile provided by another embodiment of the present application. In a specific embodiment, the fire monitoring method for a charging pile can be applied to Figure 1 The processing device 300 in the charging pile fire monitoring system shown in FIG. 3 is used as an example to Figure 4 The process shown in FIG. 1 is described in detail. The fire monitoring method for a charging pile may include the following steps 210 to 240.

[0111] Step 210: Acquire the current charging pile image of the charging pile.

[0112] Step 220: Input the current charging pile image into the target feature extraction network model to obtain a target feature map.

[0113] In this embodiment, step 210 and step 220 may refer to the contents of the corresponding steps in the aforementioned embodiment, and will not be repeated here.

[0114] Step 230: Perform feature aggregation on the target feature map to obtain an aggregated feature map.

[0115] In this embodiment, the processing device inputs the current charging pile image into the target feature extraction network model. After obtaining the target feature map, feature aggregation can be performed on the target feature map to obtain an aggregated feature map.

[0116] The aggregated feature graph may include global diversity features and global specific features.

[0117] Specifically, the processing device inputs the current charging pile image into the target feature extraction network model. After obtaining the target feature map, global average pooling and global maximum pooling can be performed on the target feature map respectively to obtain the corresponding global average pooling feature map and global maximum pooling feature map, and the global average pooling feature map and the global maximum pooling feature map are fused to obtain the aggregated feature map, such as Figure 5 shown.

[0118] Based on the global average pooling of the target feature map, a global average pooling feature map is obtained, so that the global average pooling feature map contains global diversity features. Based on the global maximum pooling of the target feature map, a global maximum pooling feature map is obtained, so that the global maximum pooling feature map contains global specific features, so that the classifier can learn the global diversity features and global specific features of the current charging pile image, which is conducive to improving the classification accuracy of the classifier.

[0119] Step 240: Input the aggregated feature map to the classifier to obtain the fire monitoring result.

[0120] In this embodiment, the processing device performs feature aggregation on the target feature map. After obtaining the aggregated feature map, the aggregated feature map can be input into the classifier. The classifier receives and responds to the aggregated feature map and outputs the fire monitoring result to the processing device. The processing device receives the fire monitoring result output by the classifier, and the classifier outputs the classification result based on the aggregated feature map. The classifier learns the global diversity features and global specific features of the current charging pile image based on the aggregated feature map, so that the classification result is more accurate, thereby making the fire monitoring result output based on the classification result more accurate, further improving the monitoring accuracy of fire monitoring on the charging pile.

[0121] The solution provided in this embodiment obtains the current charging pile image of the charging pile, inputs the current charging pile image into the target feature extraction network model, obtains the target feature map, performs feature aggregation on the target feature map, obtains the aggregated feature map, and inputs the aggregated feature map into the classifier to obtain the fire monitoring result. The classifier outputs the classification result according to the aggregated feature map. The classifier learns the global diversity features and global specific features of the current charging pile image according to the aggregated feature map, so that the classification result is more accurate, thereby making the fire monitoring result output according to the classification result more accurate, further improving the monitoring accuracy of fire monitoring on the charging pile.

[0122] See also Figure 6 , which shows a flow chart of a fire monitoring method for a charging pile provided by another embodiment of the present application. In a specific embodiment, the fire monitoring method for a charging pile can be applied to Figure 1 The processing device 300 in the charging pile fire monitoring system shown in FIG. 3 is used as an example to Figure 6 The process shown in FIG. 1 is described in detail. The fire monitoring method for a charging pile may include the following steps 310 to 360.

[0123] Step 310: Build a semantic mining model.

[0124] In this embodiment, the processing device can integrate the first Permute function module, Flatten function module, LayerNom module, Linear Encoder module, Linear Decoder module, ReLU function module, Reshape module and the second Permute function module, so that the first Permute function module, Flatten function module, LayerNom module, Linear Encoder module, Linear Decoder module, ReLU function module, Reshape module and the second Permute function module are connected in sequence to obtain a semantic mining model, thereby enhancing the semantic mining model's ability to extract global semantic features.

[0125] Step 320: Fuse the ResNet50 network model and the semantic mining model to obtain an initial feature extraction network model.

[0126] In this embodiment, after the processing device constructs the semantic mining model, the semantic mining model can be spliced ​​between the third stage module and the fourth stage module of the ResNet50 network model to obtain an initial feature extraction network model, thereby improving the extraction accuracy of the initial feature extraction network model for image features.

[0127] Among them, such as Figure 7 As shown, the initial feature extraction network model may include a two-dimensional convolution module, a first stage module, a second stage module, a third stage module, a semantic mining model and a fourth stage module connected in sequence.

[0128] Step 330: Train the initial feature extraction network model to obtain a target feature extraction network model.

[0129] In this embodiment, after the processing device fuses the ResNet50 network model and the semantic mining model to obtain an initial feature extraction network model, the initial feature extraction network model can be trained to obtain a target feature extraction network model.

[0130] In some embodiments, the processing device can obtain historical charging pile images of the charging pile, and annotate the historical charging pile images to obtain an initial sample set, and input the initial sample set into the initial feature extraction network model. The initial feature extraction network model receives and responds to the initial sample set, and is trained according to the initial sample set to obtain a target feature extraction network model.

[0131] Among them, the processing device can mark the flame information on the historical charging pile images to obtain the corresponding initial sample set.

[0132] When annotating historical charging pile images, the processing device primarily annotates the flame position and size in the historical charging pile images. For example, the processing device annotates the flame image in the historical charging pile image with an annotation box, annotates the coordinates of the corner points of the annotation box, and annotates the size of the annotation box.

[0133] In some embodiments, the processing device can obtain historical charging pile images of the charging pile, and annotate the historical charging pile images to obtain an initial sample set, and perform data enhancement processing on the initial sample set to obtain an enhanced sample set, and input the enhanced sample set into the initial feature extraction network model. The initial feature extraction network model receives and responds to the enhanced sample set, and is trained according to the enhanced sample set to obtain a target feature extraction network model, thereby realizing the training of the initial feature extraction network model according to the enhanced sample set, avoiding the low robustness of the target feature extraction network model obtained by training the initial feature extraction network model with a smaller initial sample set due to less historical charging pile image data, and increasing the robustness of the target feature extraction network model.

[0134] Among them, data enhancement processing can include at least any one of brightness enhancement processing, grayscale enhancement processing, contrast enhancement processing and transparency enhancement processing. The type of data enhancement processing is not limited here and can be set according to actual needs.

[0135] Step 340: Obtain the current charging pile image of the charging pile.

[0136] Step 350: Input the current charging pile image into the target feature extraction network model to obtain a target feature map.

[0137] Step 360: Input the target feature map into the classifier to obtain the fire monitoring result.

[0138] In this embodiment, step 340, step 350 and step 360 can refer to the contents of the corresponding steps in the above embodiments, and will not be repeated here.

[0139] The solution provided in this embodiment constructs a semantic mining model and integrates the ResNet50 network model and the semantic mining model to obtain an initial feature extraction network model, and trains the initial feature extraction network model to obtain a target feature extraction network model, and obtains the current charging pile image of the charging pile, and inputs the current charging pile image into the target feature extraction network model to obtain a target feature map, and inputs the target feature map into the classifier to obtain a fire monitoring result, thereby realizing fire monitoring of charging piles based on the target feature extraction network model and classifier constructed according to the ResNet50 network model and the semantic mining model, which can avoid missing the flames of the charging piles due to manual inspections of the charging piles, and improves the monitoring accuracy of fire monitoring of the charging piles.

[0140] The semantic mining model is used to extract and associate the global semantic features of the current charging pile image. A target feature extraction network model is constructed based on the ResNet50 network model and the semantic mining model, which improves the accuracy of the target feature extraction network model in extracting flame features, making small flames easier to detect, which is conducive to further improving the monitoring accuracy of fire monitoring of charging piles.

[0141] See also Figure 8 , which shows a fire monitoring device 500 for a charging pile provided by another embodiment of the present application. In a specific embodiment, the fire monitoring device 500 for a charging pile can be applied to Figure 1 The processing device 300 in the charging pile fire monitoring system shown in FIG. 3 is used as an example to Figure 8 The fire monitoring device 500 of the charging pile shown is described in detail. The fire monitoring device 500 of the charging pile may include an acquisition module 510 , a first input module 520 and a second input module 530 .

[0142] The acquisition module 510 can be used to obtain the current charging pile image of the charging pile; the first input module 520 can be used to input the current charging pile image into the target feature extraction network model to obtain a target feature map. The target feature extraction network model can be obtained based on the ResNet50 network model and the semantic mining model. The semantic mining model can be used to extract and associate the global semantic features of the current charging pile image; the second input module 530 can be used to input the target feature map into the classifier to obtain fire monitoring results.

[0143] In some embodiments, the fire monitoring device 500 of the charging pile may further include an aggregation module.

[0144] The aggregation module can be used for the second input module 530 to input the target feature map to the classifier, and before obtaining the fire monitoring result, perform feature aggregation on the target feature map to obtain an aggregated feature map.

[0145] In some embodiments, the second input module 530 may include an input unit.

[0146] The input unit can be used to input the aggregated feature map into the classifier to obtain the fire monitoring results.

[0147] In some embodiments, the aggregation module may include a pooling unit and a first fusion unit.

[0148] The pooling unit can be used to perform global average pooling and global maximum pooling on the target feature map respectively to obtain the corresponding global average pooling feature map and global maximum pooling feature map; the first fusion unit can be used to fuse the global average pooling feature map and the global maximum pooling feature map to obtain an aggregated feature map.

[0149] In some embodiments, the fire monitoring device 500 for a charging pile may further include a construction module, a fusion module, and a training module.

[0150] The construction module can be used to construct a semantic mining model before the acquisition module 510 obtains the current charging pile image of the charging pile; the fusion module can be used to fuse the ResNet50 network model and the semantic mining model to obtain an initial feature extraction network model; the training module can be used to train the initial feature extraction network model to obtain a target feature extraction network model.

[0151] In some embodiments, the ResNet50 network model may include a two-dimensional convolution module, a first stage module, a second stage module, a third stage module, and a fourth stage module connected in sequence, and the fusion module may include a splicing unit.

[0152] The splicing unit can be used to splice the semantic mining model between the third stage module and the fourth stage module of the ResNet50 network model to obtain the initial feature extraction network model.

[0153] In some embodiments, a building block can include a second fusion unit.

[0154] The second fusion unit can be used to fuse the first transposition module, the flattening module, the layer normalization module, the linear layer encoding module, the linear layer decoding module, the rectified linear unit activation function module, the size conversion module and the second transposition module to obtain a semantic mining model.

[0155] The first transposition module can be used to transpose the received initial feature map from the first dimension to the second dimension to obtain the first feature map; the flattening module can be used to flatten the dimension of the first feature map to obtain the second feature map; the layer normalization module can be used to perform layer normalization on the second feature map to obtain the third feature map; the linear layer encoding module can be used to perform linear transformation and mapping on the third feature map to obtain the fourth feature map; the linear layer decoding module can be used to extract and associate the global features of the fourth feature map to obtain the fifth feature map; the rectified linear unit activation function module can be used to perform rectified linear unit activation function processing on the fifth feature map to obtain the sixth feature map; the size conversion module can be used to convert the dimension of the sixth feature map to the second dimension to obtain the seventh feature map; the second transposition module can be used to transpose the seventh feature map from the second dimension to the first dimension to obtain and output the eighth feature map.

[0156] In some embodiments, the acquisition module 510 may include a detection unit and a control unit.

[0157] The detection unit can be used to perform short circuit detection and open circuit detection on the charging pile to obtain the detection results; the control unit can be used to control the camera to capture images of the charging pile when it is determined that the charging pile is in an available charging state according to the detection results, and obtain the current charging pile image.

[0158] In some embodiments, the fire monitoring device 500 for the charging pile may further include a generation module.

[0159] The generation module can be used to generate fire warning information and control the fire extinguishing equipment to extinguish the charging pile when the fire monitoring result includes the first monitoring result. The first monitoring result can be used to indicate that the current charging pile image contains flame information.

[0160] The solution provided in this embodiment obtains a current charging pile image of the charging pile and inputs the current charging pile image into a target feature extraction network model to obtain a target feature map. The target feature extraction network model is obtained based on a ResNet50 network model and a semantic mining model. The semantic mining model is used to extract and associate global semantic features of the current charging pile image, and input the target feature map into a classifier to obtain fire monitoring results. Fire monitoring of charging piles is realized by a target feature extraction network model and a classifier constructed based on the ResNet50 network model and the semantic mining model. This can avoid missing the flames of charging piles due to manual inspections of charging piles, thereby improving the accuracy of fire monitoring of charging piles.

[0161] The semantic mining model is used to extract and associate the global semantic features of the current charging pile image. A target feature extraction network model is constructed based on the ResNet50 network model and the semantic mining model, which improves the accuracy of the target feature extraction network model in extracting flame features, making small flames easier to detect, which is conducive to further improving the monitoring accuracy of fire monitoring of charging piles.

[0162] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to in detail. For the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiments. Any processing method described in the method embodiment can be implemented by the corresponding processing module in the device embodiment, and will not be repeated in detail in the device embodiment.

[0163] In addition, the functional modules in the various embodiments of the present application may be integrated into a processing module, or each module may exist physically separately, or two or more modules may be integrated into a single module. The above-mentioned integrated modules may be implemented in the form of hardware or software functional modules.

[0164] See also Figure 9 , which shows a functional block diagram of an electronic device 600 provided by an embodiment of the present application. The electronic device 600 may include one or more of the following components: a memory 610, a processor 620, and one or more applications, wherein the one or more applications may be stored in the memory 610 and configured to be executed by the one or more processors 620, and the one or more applications are configured to execute the method described in the aforementioned method embodiment.

[0165] The memory 610 may include a random access memory (RAM) or a read-only memory (ROM). The memory 610 may be used to store instructions, programs, codes, code sets, or instruction sets. The memory 610 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as obtaining a current charging pile image, inputting a current charging pile image, obtaining a target feature map, extracting global semantic features, associating global semantic features, inputting a target feature map, obtaining fire monitoring results, feature aggregation, obtaining an aggregated feature map, inputting an aggregated feature map, global average pooling, global maximum pooling, obtaining a global average pooling feature map, obtaining a global maximum pooling feature map, fusing a global average pooling feature map and a global maximum pooling feature map, constructing a semantic mining model, fusing a ResNet50 network model and a semantic mining model, obtaining an initial feature extraction network model, training an initial feature extraction network model, obtaining a target feature extraction network model, splicing a ResNet50 network model and a semantic mining model, performing short circuit detection, performing open circuit detection, obtaining detection results, determining an available charging state, controlling a camera, collecting images, obtaining an image of the current charging pile, generating fire warning information, and controlling fire extinguishing equipment, etc.), instructions for implementing the following method embodiments, etc. The storage data area can also store data created by the electronic device 600 during use (such as charging piles, current charging pile images, target feature extraction network models, target feature maps, ResNet50 network models, semantic mining models, global semantic features, classifiers, fire monitoring results, aggregated feature maps, global average pooling feature maps, global maximum pooling feature maps, initial feature extraction network models, two-dimensional convolution modules, first Stage modules, second Stage modules, third Stage modules, fourth Stage modules, first transposition modules, flattening modules, layer normalization modules, linear layer encoding modules, linear layer decoding modules, rectified linear unit activation function modules, size conversion modules, second transposition modules, initial feature maps, first dimension, second dimension, first feature maps, second feature maps, third feature maps, fourth feature maps, global features, fifth feature maps, sixth feature maps, seventh feature maps, eighth feature maps, detection results, available charging status, cameras, first monitoring results, fire warning information, fire extinguishing equipment and flame information), etc.

[0166] The processor 620 may include one or more processing cores. The processor 620 utilizes various interfaces and circuits to connect various components within the electronic device 600. It executes instructions, programs, code sets, or instruction sets stored in the memory 610, and accesses data stored in the memory 610 to perform various functions and process data within the electronic device 600. Optionally, the processor 620 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 620 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem handles wireless communications. It is understood that the modem may also be implemented independently of the processor 620 via a separate communications chip.

[0167] Please refer to Figure 10 , which shows a block diagram of a computer-readable storage medium provided in an embodiment of the present application. The computer-readable storage medium 700 stores program code 710, which can be called by a processor to execute the method described in the above method embodiment.

[0168] Computer-readable storage medium 700 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, a hard disk, or ROM. Alternatively, computer-readable storage medium 700 may include non-transitory computer-readable storage medium. Computer-readable storage medium 700 has storage space for program code 710 for executing any of the method steps described above. This program code can be read from or written to one or more computer program products. Program code 710 may be compressed, for example, in a suitable format.

[0169] Please refer to Figure 11, which shows a block diagram of the structure of a computer program product 800 provided in an embodiment of the present application. The computer program product 800 includes a computer program / instructions 810, which is stored in a computer-readable storage medium of a computer device. When the computer program product 800 is executed on the computer device, the computer device's processor reads the computer program / instructions 810 from the computer-readable storage medium and executes the computer program / instructions 810, causing the computer device to perform the method described in the above method embodiment.

[0170] The solution provided in this embodiment obtains a current charging pile image of the charging pile and inputs the current charging pile image into a target feature extraction network model to obtain a target feature map. The target feature extraction network model is obtained based on a ResNet50 network model and a semantic mining model. The semantic mining model is used to extract and associate global semantic features of the current charging pile image, and input the target feature map into a classifier to obtain fire monitoring results. Fire monitoring of charging piles is realized by a target feature extraction network model and a classifier constructed based on the ResNet50 network model and the semantic mining model. This can avoid missing the flames of charging piles due to manual inspections of charging piles, thereby improving the accuracy of fire monitoring of charging piles.

[0171] The semantic mining model is used to extract and associate the global semantic features of the current charging pile image. A target feature extraction network model is constructed based on the ResNet50 network model and the semantic mining model, which improves the accuracy of the target feature extraction network model in extracting flame features, making small flames easier to detect, which is conducive to further improving the monitoring accuracy of fire monitoring of charging piles.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A fire monitoring method for a charging pile, characterized in that: include: Get the current charging pile image of the charging pile; Inputting the current charging pile image into a target feature extraction network model to obtain a target feature map, wherein the target feature extraction network model is obtained based on a ResNet50 network model and a semantic mining model, and the semantic mining model is used to extract and associate global semantic features of the current charging pile image; Inputting the target feature map into the classifier to obtain the fire monitoring result; Before obtaining the current charging pile image of the charging pile, the fire monitoring method further includes: Constructing the semantic mining model; The ResNet50 network model and the semantic mining model are integrated to obtain an initial feature extraction network model; Training the initial feature extraction network model to obtain the target feature extraction network model; The ResNet50 network model includes a two-dimensional convolution module, a first stage module, a second stage module, a third stage module, and a fourth stage module connected in sequence. The initial feature extraction network model is obtained by fusing the ResNet50 network model and the semantic mining model, including: Splicing the semantic mining model between the third stage module and the fourth stage module of the ResNet50 network model to obtain the initial feature extraction network model; wherein the third stage module includes 6 bottleneck layers, and the fourth stage module includes 3 bottleneck layers; The constructing of the semantic mining model includes: fusing a first transposition module, a flattening module, a layer normalization module, a linear layer encoding module, a linear layer decoding module, a rectified linear unit activation function module, a size conversion module, and a second transposition module to obtain the semantic mining model; The first transposition module is used to transpose the received initial feature map from the first dimension to the second dimension to obtain a first feature map; the flattening module is used to flatten the dimension of the first feature map to obtain a second feature map; the layer normalization module is used to perform layer normalization processing on the second feature map to obtain a third feature map; the linear layer encoding module is used to perform linear transformation and mapping on the third feature map to obtain a fourth feature map; the linear layer decoding module is used to extract and associate global features of the fourth feature map to obtain a fifth feature map; the rectified linear unit activation function module is used to perform rectified linear unit activation function processing on the fifth feature map to obtain a sixth feature map; the size conversion module is used to convert the dimension of the sixth feature map to the second dimension to obtain a seventh feature map; the second transposition module is used to transpose the seventh feature map from the second dimension to the first dimension to obtain and output an eighth feature map; The fire monitoring method is used to detect small flames.

2. The fire monitoring method according to claim 1, characterized in that: Before inputting the target feature map into the classifier to obtain the fire monitoring result, the fire monitoring method further includes: Performing feature aggregation on the target feature map to obtain an aggregated feature map; The target feature map is input to the classifier to obtain a fire monitoring result; The aggregated feature map is input into the classifier to obtain the fire monitoring result.

3. The fire monitoring method according to claim 2, characterized in that: The performing feature aggregation on the target feature map to obtain an aggregated feature map includes: Performing global average pooling and global maximum pooling on the target feature map respectively to obtain corresponding global average pooling feature map and global maximum pooling feature map; The global average pooling feature map and the global maximum pooling feature map are fused to obtain the aggregated feature map.

4. The fire monitoring method according to any one of claims 1 to 3, characterized in that: The obtaining of the current charging pile image of the charging pile includes: Performing short circuit detection and open circuit detection on the charging pile to obtain detection results; When it is determined according to the detection result that the charging pile is in an available charging state, the camera is controlled to capture an image of the charging pile to obtain the current charging pile image.

5. The fire monitoring method according to any one of claims 1 to 3, characterized in that: Also includes: When the fire monitoring result includes a first monitoring result, fire warning information is generated, and a fire extinguishing device is controlled to extinguish the fire on the charging pile. The first monitoring result is used to indicate that the current charging pile image contains flame information.

6. An electronic device, characterized in that: include: Memory; one or more processors coupled to the memory; One or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, and the one or more application programs are configured to execute the fire monitoring method according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program codes, which can be called by a processor to execute the fire monitoring method according to any one of claims 1 to 5.

Images