Transformer fire detection method based on infrared and visible light image fusion

Abstract

The invention provides a transformer fire detection method based on infrared and visible light image fusion, and the system comprises an infrared camera, a visible light camera, an image fusion and analysis host, a large monitoring screen, and an alarm device. The use method of the detection system comprises the steps that infrared images and visible light images captured in real time are transferred into the image fusion and analysis host through the infrared camera and the visible light camera, and after the image fusion and analysis host receives the image flow, the infrared images and the visible light images are fused through an image fusion algorithm; a multi-modal image is formed, then a pre-trained deep learning model is utilized to analyze the fused image, an early occurrence sign of a fire disaster is identified, a fire hazard part is found in time, after identification, the fire hazard part is presented to a large monitoring screen, and meanwhile, a sound-light alarm device is triggered so as to timely remind related operators on duty to rapidly take measures. Early-stage fire disasters are extinguished, and operation safety of the transformer is guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of image monitoring, and in particular to a transformer fire detection method based on infrared and visible light image fusion. Background Art

[0002] Transformers are important equipment in power systems. Once a fire occurs, it will not only cause damage to the equipment, but may also cause large-scale power outages and even endanger personnel safety. Therefore, timely and accurate detection of transformer fires and prevention is crucial for the safe operation of power systems. However, traditional transformer fire detection methods mainly rely on visible light images, which have limitations in some cases, especially in the early stages of a fire, when the flames are small and not easily recognized by visible light sensors. The accuracy and timeliness of detection may be insufficient. For example: A lightweight convolutional neural network fire detection method, device and equipment, application publication number: CN115424116A; the method includes obtaining an image of a UHV converter transformer area; inputting the image of the UHV converter transformer area into a prediction First, train the fire recognition network model to obtain the fire detection results; the fire recognition network model includes the SSD target recognition network, the backbone feature extraction network of the SSD target recognition network is the lightweight network MobileNetV2, and the attention mechanism Squeeze-and-Excitation is added to the lightweight network MobileNetV2; based on the lightweight network MobilNetV2, the SE attention mechanism is added to achieve a faster running speed with minimal increase in computational complexity without significant loss of accuracy; although by introducing the lightweight network MobileNetV2 and the SE attention mechanism, faster fire detection can be achieved with a smaller amount of computation, which is especially suitable for the processing of UHV converter transformer area images. However, this technology mainly relies on the analysis of visible light images. For situations where the flames are small in the early stage of a fire or due to interference from ambient light, detection that relies solely on visible light images may have the following shortcomings:

[0003] 1. Detection lag: In the early stages of a fire, when the flames are small or hidden inside the equipment, visible light images may not be able to effectively capture early fire signals, resulting in detection lag;

[0004] 2. Ambient light interference: Environmental factors such as strong light or shadows may affect the quality of visible light images, thereby affecting the accuracy of detection results;

[0005] 3. Limited temperature perception: Visible light images cannot provide accurate temperature distribution information. When flames cannot be directly observed, it is impossible to warn of fires in advance through temperature anomalies.

[0006] Today, infrared imaging technology is very mature. By capturing the thermal radiation of objects, flames or high-temperature areas can be effectively detected. Even in the early stages of a fire when the flames are small or in poor light conditions, infrared images can still provide clear temperature distribution information. To this end, the applicant proposes a transformer fire detection method based on the fusion of infrared and visible light images according to the needs of transformer fire monitoring. Combining infrared images with visible light images can greatly improve the reliability and accuracy of fire detection, and has stronger early detection capabilities. It can still work effectively under environmental interference such as strong light and shadows, and is not affected by the limitations of visible light images, ensuring the stability of detection. Summary of the invention

[0007] In order to solve the above technical problems, the present invention proposes a transformer fire detection method based on infrared and visible light image fusion. By setting an infrared camera, a visible light camera, an image fusion and analysis host, a monitoring screen and an alarm device, the detection system is used in that the infrared image and visible light image captured in real time are transmitted to the image fusion and analysis host through the infrared camera and the visible light camera. After the image fusion and analysis host receives the image stream, the infrared image and the visible light image are fused through the image fusion algorithm to form a multimodal image, and then the fused image is analyzed using a pre-trained deep learning model to identify the early signs of fire, and the fire hazard sites are discovered in time. After identification, they are presented to the monitoring screen and the sound and light alarm device is triggered at the same time, so as to promptly remind the relevant on-duty personnel to take prompt measures to extinguish the early fire and ensure the safe operation of the transformer.

[0008] To achieve the above object, the technical solution adopted by the present invention is:

[0009] A transformer fire detection method based on infrared and visible light image fusion, the specific steps of the method are as follows:

[0010] Step 1: Infrared cameras and visible light cameras collect transformer operation video images in real time and transmit them to the image fusion and analysis host through the RTSP protocol. The protocol supports real-time streaming transmission, ensuring the timeliness and continuity of the images.

[0011] Step 2: In the image fusion and analysis host, the received infrared image and visible light image are first preprocessed, including denoising, alignment and scale adjustment. Then, the image fusion algorithm is used to fuse the temperature information of the infrared image with the visual information of the visible light image to generate a multimodal image containing the two types of information.

[0012] Step 3: The fused multimodal image is analyzed by a fire detection model; the model uses a pre-trained convolutional neural network that can identify flames, smoke, and abnormal temperature areas in the image; if early signs of a fire are detected, such as flames or a significant temperature increase, the model will generate a fire alarm signal;

[0013] Step 4: Image fusion and analysis The host transmits the processed image and detection results to the monitoring screen in real time, and the on-duty personnel can monitor the status of the transformer through the screen;

[0014] Step 5: When a fire is detected, the system will display an alarm message on the monitoring screen and trigger an external alarm device. At the same time, the system can be set to send early warning information to the mobile phones or other devices of relevant personnel to ensure that the fire situation can be handled in the first time;

[0015] Step 6: To ensure the reliability of the system, analyze the host and set up a redundant backup system to prevent the detection system from failing due to hardware failure;

[0016] Step 7: After the transformer fire hazard is eliminated or the on-duty personnel have checked on site and confirmed that there is no fire and no action is required, perform the reset operation and the alarm information will be cleared.

[0017] Furthermore, the image fusion and analysis host of the transformer fire detection system and method based on infrared and visible light image fusion adopts a dual-channel information fusion method, fully utilizing the complementary advantages of the two in fire monitoring, and proposes a bidirectional structure based on the YOLOv5 network, which is used for infrared and visible light target detection networks respectively, to solve the problem of insufficient texture features of a single infrared image in target detection, and to improve the overall detection accuracy by supplementing the feature information of infrared and visible light images.

[0018] Furthermore, the transformer fire detection system and method based on infrared and visible light image fusion supports the access of multiple cameras, is suitable for the monitoring needs of large-scale substations, and the system's detection algorithm can be customized and optimized according to specific environmental requirements.

[0019] Furthermore, the transformer fire detection system and method based on infrared and visible light image fusion adopts infrared imaging technology, which can achieve early warning through temperature anomalies in the early stage of fire, that is, when the flames are small or not yet visible.

[0020] Furthermore, the transformer fire detection system and method based on infrared and visible light image fusion can still perform effective fire detection in strong light, shadow or at night through the temperature perception ability of infrared images, and is not affected by the limitations of visible light images.

[0021] Furthermore, the transformer fire detection system includes an infrared camera, a visible light camera, an image fusion and analysis host, a monitoring screen and an alarm device; the infrared camera and the visible light camera are installed in an area where it is convenient to monitor the transformer operation and where fire is prone to occur, and the image fusion and analysis host, the monitoring screen and the alarm information are installed in the duty room.

[0022] The benefits of this application are:

[0023] 1. Transformer fire detection system and method based on infrared and visible light image fusion combines the advantages of infrared images and visible light images. The fusion of multimodal images greatly improves the reliability and accuracy of fire detection;

[0024] 2. The transformer fire detection system and method based on infrared and visible light image fusion adopts infrared imaging technology, which can warn of fire in advance by detecting temperature anomalies when the fire has just occurred and the flames are not yet clearly visible. The early detection capability significantly improves the effectiveness of transformer fire protection and reduces the risk of fire spread and equipment damage;

[0025] 3. Transformer fire detection system and method based on infrared and visible light image fusion. By combining the temperature perception capability of infrared images, the system can still work effectively under environmental interference such as strong light and shadows, and is not affected by the limitations of visible light images, ensuring the stability of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Schematic diagram of the bidirectional structure target detection network structure of the system of the present invention. DETAILED DESCRIPTION

[0027] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments:

[0028] like Figure 1 As shown, a transformer fire detection method based on infrared and visible light image fusion is shown. The transformer fire detection system based on infrared and visible light image fusion includes an infrared camera, a visible light camera, an image fusion and analysis host, a monitoring screen and an alarm device; the infrared camera and the visible light camera are installed in an area where it is convenient to monitor the transformer operation and prone to fire, and the image fusion and analysis host, the monitoring screen and the alarm information are installed in the duty room; the specific steps of the method of using the detection system are as follows:

[0029] Step 1: Infrared cameras and visible light cameras collect real-time video images of transformer operation and transmit them to the image fusion and analysis host through the RTSP protocol. The protocol supports real-time streaming transmission, ensuring the timeliness and continuity of the images. As shown, infrared cameras and visible light cameras are installed near each transformer in the substation, and the cameras are aimed at the transformer to capture infrared images and visible light images of the transformer in real time. The infrared camera is used to capture the temperature distribution image of the transformer, and the visible light camera is used to capture the visual image of the transformer.

[0030] Step 2: In the image fusion and analysis host, the received infrared image and visible light image are first preprocessed, including denoising, alignment and scale adjustment. Then, the image fusion algorithm is used to fuse the temperature information of the infrared image with the visual information of the visible light image to generate a multimodal image containing the two types of information.

[0031] Step 3: The fused multimodal image is analyzed by a fire detection model; the model uses a pre-trained convolutional neural network that can identify flames, smoke, and abnormal temperature areas in the image; if early signs of a fire are detected, such as flames or a significant temperature increase, the model will generate a fire alarm signal;

[0032] Step 4: Image fusion and analysis The host transmits the processed image and detection results to the monitoring screen in real time, and the on-duty personnel can monitor the status of the transformer through the screen;

[0033] Step 5: When a fire is detected, the system will display an alarm message on the monitoring screen and trigger an external alarm device. At the same time, the system can be set to send early warning information to the mobile phones or other devices of relevant personnel to ensure that the fire situation can be handled in the first time;

[0034] Step 6: To ensure the reliability of the system, analyze the host and set up a redundant backup system to prevent the detection system from failing due to hardware failure;

[0035] Step 7: After the transformer fire hazard is eliminated or the on-duty personnel have checked on site and confirmed that there is no fire and no action is required, perform the reset operation and the alarm information will be cleared.

[0036] The image fusion and analysis host of the transformer fire detection system and method based on infrared and visible light image fusion shown in the figure adopts a dual-channel information fusion method, fully utilizing the complementary advantages of the two in fire monitoring. On the basis of the YOLOv5 network, a bidirectional structure is proposed, which is used for infrared and visible light target detection networks respectively, to solve the problem of insufficient texture features of a single infrared image in target detection, and to improve the overall detection accuracy by supplementing the feature information of infrared and visible light images.

[0037] The transformer fire detection system and method based on infrared and visible light image fusion supports the access of multiple cameras and is suitable for the monitoring needs of large-scale substations. The detection algorithm of the system can be customized and optimized according to specific environmental requirements.

[0038] The transformer fire detection system and method based on infrared and visible light image fusion shown adopts infrared imaging technology, which can achieve early warning through temperature abnormalities in the early stage of fire, that is, when the flames are small or not yet visible.

[0039] The transformer fire detection system and method based on infrared and visible light image fusion can still perform effective fire detection in strong light, shadow or at night through the temperature perception ability of infrared images, and is not affected by the limitations of visible light images.

[0040] The above description is only a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent change made based on the technical essence of the present invention still falls within the scope of protection required by the present invention.

Claims

1. A transformer fire detection method based on infrared and visible light image fusion, characterized in that: The specific steps for use are: Step 1: Infrared cameras and visible light cameras collect transformer operation video images in real time and transmit them to the image fusion and analysis host through the RTSP protocol. The protocol supports real-time streaming transmission, ensuring the timeliness and continuity of the images. Step 2: In the image fusion and analysis host, the received infrared image and visible light image are first preprocessed, including denoising, alignment and scale adjustment. Then, the image fusion algorithm is used to fuse the temperature information of the infrared image with the visual information of the visible light image to generate a multimodal image containing the two types of information. Step 3: The fused multimodal image is analyzed by a fire detection model; the model uses a pre-trained convolutional neural network that can identify flames, smoke, and abnormal temperature areas in the image; if early signs of a fire are detected, such as flames or a significant temperature increase, the model will generate a fire alarm signal; Step 4: Image fusion and analysis The host transmits the processed image and detection results to the monitoring screen in real time, and the on-duty personnel can monitor the status of the transformer through the screen; Step 5: When a fire is detected, the system will display an alarm message on the monitoring screen and trigger an external alarm device. At the same time, the system can be set to send early warning information to the mobile phones or other devices of relevant personnel to ensure that the fire situation can be handled in the first time; Step 6: To ensure the reliability of the system, analyze the host and set up a redundant backup system to prevent the detection system from failing due to hardware failure; Step 7: After the transformer fire hazard is eliminated or the on-duty personnel have checked on site and confirmed that there is no fire and no action is required, perform the reset operation and the alarm information will be cleared.

2. The transformer fire detection method based on infrared and visible light image fusion according to claim 1 is characterized in that: The image fusion and analysis host of the transformer fire detection system and method based on infrared and visible light image fusion adopts a dual-channel information fusion method, fully utilizing the complementary advantages of the two in fire monitoring. On the basis of the YOLOv5 network, a bidirectional structure is proposed, which is used for infrared and visible light target detection networks respectively, to solve the problem of insufficient texture features of a single infrared image in target detection, and to improve the overall detection accuracy by supplementing the feature information of infrared and visible light images.

3. The transformer fire detection method based on infrared and visible light image fusion according to claim 1 is characterized in that: The transformer fire detection system and method based on infrared and visible light image fusion supports the access of multiple cameras, is suitable for the monitoring needs of large-scale substations, and the detection algorithm of the system can be customized and optimized according to specific environmental requirements.

4. A transformer fire detection method based on infrared and visible light image fusion according to claim 1, characterized in that: The transformer fire detection system and method based on infrared and visible light image fusion adopts infrared imaging technology, and can achieve early warning through temperature anomalies in the early stage of fire, that is, when the flames are small or not yet visible.

5. A transformer fire detection method based on infrared and visible light image fusion according to claim 1, characterized in that: The transformer fire detection system and method based on infrared and visible light image fusion can still perform effective fire detection in strong light, shadow or at night through the temperature perception ability of infrared images, and is not affected by the limitations of visible light images.

6. A transformer fire detection method based on infrared and visible light image fusion according to claim 1, characterized in that: The transformer fire detection system includes an infrared camera, a visible light camera, an image fusion and analysis host, a monitoring screen and an alarm device; the infrared camera and the visible light camera are installed in an area where it is convenient to monitor the transformer operation and prone to fire, and the image fusion and analysis host, the monitoring screen and the alarm information are installed in the duty room.

Images